Skip to main content

Full text of "Year book - Carnegie Institution of Washington"

See other formats


P w A :< H I v ; 



J^^ /),*■'**>*"*' °ft" 



Year Book 74 


Library of Congress Catalog Card Number 3-16716 

The John D. Lucas Printing Company, Baltimore, Maryland 

Issued December 1975 



Officers and Staff v 

Report of the President 1 

Reports of Departments and Special Studies 1 

Department of Embryology 3 

Department of Terrestrial Magnetism 105 

Hale Observatories 303 

Geophysical Laboratory 379 

Developmental Biology Research Group 651 

Department of Plant Biology 737 

Bibliography -815 

Administrative Reports 817 

Report of the Executive Committee 819 

Abstract of Minutes of the Seventy-Sixth Meeting of the 

Board of Trustees 821 

Financial Statement 823 

Ten-Year Financial Summary 825 

Report of Independent Public Accountants 826 

Articles of Incorporation 841 

By-Laws of the Institution 845 

Index of Names 851 


Digitized by the Internet Archive 

in 2012 with funding from 

LYRASIS Members and Sloan Foundation 

President and Trustees 


Philip H. Abelson 


William McChesney Martin, Jr. 

Frank Stanton 

William T. Golden 

Eric Ashby 1 
Lewis M. Branscomb 
John T. Connor 2 
John Diebold 2 
Michael Ference, Jr. 
Carl J. Gilbert 
Hanna H. Gray 
Crawford H. Greenewalt 
William C. Greenough 2 
Patrick E. Haggerty 3 
Caryl P. Haskins 
William R. Hewlett 
Keith S. McHugh 4 
Henry S. Morgan 
William I. Myers 
Walter H. Page 
Robert M. Pennoyer 
Richard S. Perkins 
William M. Roth 
Robert C. Seamans, Jr. 
Charles P. Taft 5 
Charles H. Townes 
Juan T. Trippe 
James N. White 

Alfred L. Loomis 6 
Garrison Norton 
Trustees Emeriti 

1 Resigned June 1974. 

2 Elected May 1975. 

3 Resigned June 1975. 

4 Resigned September 1974; designated Trustee Emeritus May 1975; died June 7, 1975. 

5 Resigned April 1975; designated Trustee Emeritus May 1975. 

6 Died August 11, 1975. 

Former Presidents and Trustees 

Daniel Coit Oilman 
John Campbell Merriam 
President Emeritus 

Alexander Agassiz 
George J. Baldwin 
Thomas Barbour 
James F. Bell 
John S. Billings 
Robert Woods Bliss 
Amory H. Bradford 
Lindsay Bradford 
Omar N. Bradley 
Robert S. Brookings 
Vannevar Bush 
John L. Cadwalader 
William W. Campbell 
John J. Carty 
Whitefoord R. Cole 
Frederic A. Delano 
Cleveland H. Dodge 
William E. Dodge 
Charles P. Fenner 
Homer L. Ferguson 
Simon Flexner 
W. Cameron Forbes 
James Forrestal 
William N. Frew 
Lyman J. Gage 
Walter S. Gifford 
Cass Gilbert 
Frederick H. Gillett 
Daniel C. Gilman 
John Hay 

Barklie McKee Henry 
Myron T. Her rick 
Abram S. Hewitt 
Henry L. Higginson 
Ethan A. Hitchcock 
Henry Hitchcock 
Herbert Hoover 
William Wirt Howe 
Charles L. Hutchinson 
Walter A. Jessup 
Frank B. Jewett 
Samuel P. Langley 
Ernest 0. Lawrence 
Charles A. Lindbergh 
William Lindsay 
Henry Cabot Lodge 


1902-1904 Vannevar Bush 1939-1955 

1921-1938 Robert Simpson Woodward 1904-1920 

1939-1945 Caryl P. Haskins 1956-1971 


Carroll D. Wright 

Robert A. Lovett 1948-71 

Seth Low 1902-16 

Wayne MacVeagh 1902-07 

Andrew W. Mellon 1924-37 

John Campbell Merriam 1921-38 

Margaret Carnegie Miller 1955-67 

Roswell Miller 1933-55 

Darius O. Mills 1902-09 

S. Weir Mitchell 1902-14 

Andrew J. Montague 1907-35 

William W. Morrow 1902-29 

Seeley G. Mudd 1940-68 

William Church Osborn 1927-34 

James Parmelee 1917-31 

Wm. Barclay Parsons 1907-32 

Stewart Paton 1916-42 

George W. Pepper 1914-19 

John J. Pershing 1930-43 

Henning W. Prentis, Jr. 1942-59 

Henry S. Pritchett 1906-36 

Gordon S. Rentschler 1946-48 

David Rockefeller 1952-56 

Elihu Root 1902-37 

Elihu Root, Jr. 1937-67 

Julius Rosen wald 1929-31 

William W. Rubey 1962-74 

Martin A. Ryerson 1908-28 

Henry R. Shepley 1937-62 

Theobald Smith 1914-34 

John C. Spooner 1902-07 

William Benson Storey 1924-39 

Richard P. Strong 1934-48 

William H. Taft 1906-15 

William S. Thayer 1929-32 

James W. Wadsworth 1932-52 

Charles D. Walcott 1902-27 

Frederic C. Walcott 1931-48 

Henry P. Walcott 1910-24 

Lewis H. Weed 1935-52 

William H. Welch 1906-34 

Andrew D. White 1902-16 

Edward D. White 1902-03 

Henry White 1913-27 

George W. Wickersham 1909-36 

Robert E. Wilson 1953-64 

Robert S. Woodward 1905-24 

Under the original charter, from the date of organization until April 28, 1904, the following 
were ex officio members of the Board of Trustees : the President of the United States, the Presi- 
dent of the Senate, the Speaker of the House of Representatives, the Secretary of the 
Smithsonian Institution, and the President of the National Academy of Sciences. 



1530 P Street, N.W., Washington, D.C. 20005 

Philip H. Abelson President 

James W. Boise Bursar ; Secretary-Treasurer , Retirement Trust; 
Executive Secretary to the Finance Committee 

Marjorie H. Walburn Assistant to the President 

Sheila A. McGough Publications Officer; Editor 

Kenneth R. Henard Assistant Bursar; Assista?it Treasurer, 

Retirement Trust 

Joseph M. S. Haraburda Assistant to the Bursar 

Richard E. Hewitt Administrative Officer for Services 

Marshall Hornblower Counsel 


Roy J. Britten 
Alfred D. Hershey 
Tatiana Proskouriakoff 


Barbara McClintock 


Harry E.D. Pollock 


Report of 
the President 

Many peoples of the world are beginning to learn the hard way what a 
few observers have known for a long time — exponential growth of any kind 
cannot go on indefinitely. For more than a hundred years much of the world has 
experienced an unprecedented era of fast growth in the use of energy and raw 
materials. In addition, since World War II exponential growth of population has 
been ubiquitous, with especially high rates prevailing in less developed countries. 
The inevitable cessation of rapid growth of population will not occur suddenly or 
universally but rather on a country by country basis. In contrast, changes in the 
world situation with respect to energy in the form of hydrocarbons have been 

For more than two decades, until the end of 1973, the use of petroleum in the 
United States, Japan, and Western Europe increased rapidly. In the developed 
countries hydrocarbons became the principal source of energy, and many observers 
have pointed to a worldwide correlation between consumption of energy and 
standard of living. In the United States, the annual rate of increase was 4%-5%. 
In the other highly developed countries, the rate was even faster. These countries 
enjoyed great prosperity. During the past two years, consumption of energy in 
all these places has decreased. Some economists argue that increased costs of oil 
represent only about 2% of the gross national product and hence should not 
affect the economy much. Perhaps so. But maybe the effects reverberate through 
the economy in ways that are not altogether understood, for example, people 
turning away from 2-ton automobiles. 

Perhaps there is a lesson to be learned from the lowly bacterium Escherichia 
coli. When a simple solution containing a food — glucose — and some inorganic 
salts is inoculated with E. coli, the bacteria enjoy a period of exponential growth. 
Starting with, for example, a population of 10 3 per milliliter, growth proceeds 
with a doubling time of 53 minutes. That is, after such an interval, the population 
increases to 2 X 10 3 per milliliter, and after a like period, to 4 X 10 3 . This doubling 
goes on with clocklike precision for a period of about 18 hours — more than 20 
doubling times during which the population increases more than a millionfold. At 
that point, if bacteria had brains, traditions, written history, and econometricians, 
they would confidently expect a continuation of a society geared to exponential 
growth. However, in the course of only a fraction of a doubling time, exponential 
growth stops and the bacterial culture goes into a static state followed by a 
decline. This, despite the fact that almost all conditions for growth, such as food 
supply, are adequate. There is little to be gained from pursuing analogies, but 
one point should be emphasized. In any biological system, including those in- 
volving humans, exponential growth can stop suddenly, to be followed by a 
declining state. 

Exponential economic growth in the United States may be restored. But it 
will not be restored to the precise conditions existing in 1973. The country faces 
a decade or more of adjustments to a new set of realities, both internal and 
external. The world situation is especially complex in that limiting factors are 
becoming apparent in energy, food, and raw materials. Moreover, constraints 
among these major factors are interactive. The key factor is energy. Given 
sufficient and inexpensive quantities of it, modern technology could provide ample 
food and materials. However, the world faces major problems as economies based 
on oil change to those utilizing other sources of energy. 

One by one the oil producers of the rest of the world will repeat the experience 



of the United States. After many years of exponential growth, domestic production 
reached a peak in 1970. During 1974 and 1975, production declined at the rate 
of 5% per year. This occurred despite a quadrupling of the price for new oil. 
Energy problems of the United States have been compounded by falling supplies 
of natural gas, the leading domestic source of energy. Production of this fuel at 
one time also followed an exponential growth curve, later tapering off and 
declining in 1974 and 1975, at the annual rate of 6%-8%. 

Prospects are poor for an overall improvement in these trends during the next 
five years. Oil is scheduled to flow in the Alaska pipeline in late 1977, but this 
will arrest only briefly the decay in national production. Some analysts believe 
optimistically that higher prices and more drilling activity and enhanced tertiary 
recovery will bring out additional supplies, but the record of the past two years 
provides no basis for such optimism. Similar remarks apply to prospects for 
natural gas during the next five years. Thus, the most realistic estimate is that 
present declining trends will continue. On this basis, in 1980 one would expect 
a domestic production of 7 million barrels of oil a day and 13 trillion cubic feet 
of gas a day. In 1974, the corresponding figures were 8.8 million and 21 trillion, 

Taken together, oil and gas supplied 75.7% of United States energy con- 
sumption in 1974. Thus simultaneous failure of the two sources is a serious 
problem. It is not practical to meet more than a small fraction of the shortfall 
of natural gas by imports. Fair prospects exist for an expansion of domestic 
supplies of natural gas after a period of perhaps five years. However, in the 
meantime, if energy consumption is to be maintained, let alone increased, the 
nation must turn to other sources. Of these, by far the most convenient is oil. 
Thus, considerable pressure is likely to build to expand greatly imports of oil. 
But this would entail increased vulnerability and balance of payments problems. 
Experience in 1973-1974 demonstrated that the United States is vulnerable when 
its imports exceed 20% of consumption. In late 1975, they were 38%, which 
amounted to about 6 million barrels a day. Merely to maintain the present level 
of consumption of energy based on hydrocarbons would require an annual 
expansion of imports of about 1 million barrels of oil a day. 

Oil and natural gas are crucial to most societal needs and to the functioning 
of the economy, e.g., merely to name a few uses — petroleum products for cars, 
buses, trains, and planes, and for agriculture. Natural gas is a feedstock for 
fertilizers and petrochemicals. It is the main source of energy for industry, and 
it supplies the heat needed by 55% of the nation's homes. Hydrocarbons are 
so essential that it is questionable whether the economy can grow when supplies 
of these fuels are static or contracting. Ultimately, the country must shift to 
other energy sources, such as coal, oil shale, wood, and solar and nuclear power. 
Such shifts will have profound effects on the economy and on the shape of society, 
and will not be accomplished without various kinds of trauma. 

The magnitude of the changes is not easy to visualize from the vantage point 
of the present. After all, houses are warm and there is gasoline at the pumps. 
However, perspective can be gained by surveying briefly the past century or so. 
Just a little over a hundred years ago the nation's principal source of energy was 
wood. It was used in homes and in locomotives. The principal source of energy 
for factories was water power, and so most of the manufacturing was located 
near streams such as those in New England. Most of the population lived on 
farms. About a hundred years ago a major shift toward coal occurred. This led 
to a tremendous industrial boom in Pittsburgh and the Middle West — a develop- 
ment in which Andrew Carnegie was a leader. This was soon accompanied by 


production and use of electricity. An important offshoot was the electric streetcar. 
As cities grew, their pattern was influenced by the new form of transportation. 

By 1920 a new phenomenon was well under way — the creation of a mobile 
society based on the internal combustion engine and on petroleum. Tractors and 
other machinery revolutionized agriculture. At one time, more than 90% of 
the population were engaged in farming. Today, only about 4% are so employed. 
Such a shift had inevitable social consequences. 

Around 1950 another important development occurred — a great expansion in 
the use of natural gas. Much of this fuel is obtained as a by-product of petroleum 
production. Originally the gas was burned as a nuisance. Then its value was 
perceived and exploited. At first the price was very small. Natural gas was much 
cheaper and, of course, cleaner than coal and more convenient to use. In conse- 
quence, total consumption of coal in the United States remained about the same 
for 25 years while other factors of the economy expanded and prospered. As a 
corollary, Appalachia endured more than two decades of stagnation. At the 
same time, Houston and the Gulf Coast prospered and the region became the 
locale of great petrochemical complexes based on cheap energy and cheap 

During the past two years there has been a fundamental change in the realities 
governing energy. Not so many years ago foreign residual oil was delivered to 
the East Coast for $1.40 a barrel (42 gallons). This was equivalent in cost to 
$0.23 per million BTU. At that time, crude oil in Texas fetched $3.00 a barrel. 
Appalachian coal at the mine mouth brought from $4 to $8 a ton, i.e., from $0.17 
to $0.34 per million BTU. Today, foreign residual oil is priced at about $10 a 
barrel. Coal as an energy source is much cheaper than oil or natural gas, and 
Appalachia is booming. New England, which is poorly endowed with indigenous 
energy, is experiencing depression and high unemployment. The Gulf Coast 
continues to prosper because of high returns on its petroleum and natural gas. 
However, its earlier marked advantages of cheap energy and feedstocks have 
disappeared. It is as yet too early to identify the next great centers of industrial 
expansion. However, it is certain that society and the economy will undergo 
transformations of a magnitude comparable to those that occurred earlier in 
this century. 

Much depends on the choices that are made. For example, how large a role 
will the various sources fill? Can suitable replacements for the hydrocarbons 
be produced? Will it be possible to maintain a mobile society? In principle, the 
answers to the last two questions are yes. But in practice, the costs may require 
considerable modifications in the use of mobile machinery. Likely sources of the 
energy are coal, shale, or wood. If liquids from coal were used, the fluid would 
not be burned in a gasoline engine but in something resembling a diesel. To 
supply such fuel, chemical complexes costing on the order of $100 billion would 
be required. If the choice were shale oil, an enormous and not altogether welcome 
concentration of facilities would be centered in Colorado and Utah. If fluids 
derived from trees or plants were employed, they would likely take the form of 
ordinary alcohol. In that case, a motor not too different from that of the present 
might be employed, though one might expect substantial improvement on it. 
If trees were used as the basic source of energy, the southeastern and north- 
western parts of the country would be especially favored. 

Another imponderable is the role of nuclear power, including the breeder. 
Of late there has been increasing resistance within the United States to expansion 
of nuclear power even though this is at present the cheapest method of generating 


electricity. In principle, nuclear plants could be located in virtually every state, 
and costs would not be much different in the various regions. Nuclear energy 
shares with several other forms, such as geothermal and wind, the handicap that 
the product is in the form of electricity. Electricity has many useful character- 
istics, but it fills directly only a small fraction of the total energy needs. If liquid 
fuels were made by a process involvingelectricity, they would be very costly. 

One can multiply examples of the consequences of decisions about energy 
within the United States. These effects will be profound. Capital expenditures 
for energy-related plants will be enormous, but they are only part of the capital 
costs. Creation of new facilities in new localities will involve concomitant con- 
struction of housing and all that entails in the way of supporting facilities such 
as roads and schools. Many older communities will be impoverished as factories 
move and young people leave. Given a prosperous United States with a growing 
economy such changes could be managed. However, prospects for a growing 
economy during the next decade or more are not good. During that time, domestic 
production of oil and natural gas will continue to decline. Attrition of supplies 
of natural gas will cause great difficulties for industry, which derives more than 
half its energy from this source. Thus the United States faces a difficult period of 
attempting to live with dislocations due to failure of supplies of hydrocarbons, 
attendant unemployment, and huge costs of welfare. 

A period of heavy federal deficit spending is hardly a favorable time to engage 
in a huge program of private capital expenditures. The paralyzing effects of 
such a situation were felt in 1975. High interest rates drove up the total costs 
of construction of new energy facilities. (In some instances, the time required to 
go on-stream is close to ten years.) In consequence, many energy projects were 
delayed or cancelled, including nuclear, shale oil, and gasification plants. They 
all would have helped ease forthcoming shortages of hydrocarbons, and the 
shale and gasification plants would also have supplied invaluable new practical 
learning experiences. 

The outlook is not all dismal. At least some parts of industry are moving to 
adapt to the evolving situation. Many companies have found ways to cut con- 
sumption of energy while maintaining production. During the past four years 
one major organization has reduced its consumption per unit product by 27%. 
Another important development is the trend toward lighter, less fuel-consuming 
automobiles. As the current inventory of more than 100 million automobiles 
is slowly replaced, demand for petroleum will be substantially reduced. 

Another aspect in which there has been some progress is the potential utiliza- 
tion of coal. This fuel could in principle replace much of the hydrocarbons that 
are burned to produce heat for boilers and for industrial processing. During 
1975 there were many announcements of plans to open new mines, thus leading 
toward increasing production. At the same time, ability to use coal was being 
improved. An important environmental constraint is the high sulfur content of 
many eastern coals, but this year a number of processes were developed to 
ameliorate this problem. One of them was an intriguing method combining 
standard coal washing with a following separation of coal into two fractions, 
one very low in sulfur and the other containing a moderate amount. The very 
low fraction can be burned under utility boilers and meet clean air standards for 
new plants. The other fraction can be burned in the older plants or in plants 
having flue gas desulfurization equipment. Separation into two fractions is done 
through the use of a magnetite-water suspension with specific gravity delicately 
adjusted to cause one fraction to float and the other to sink. 

In its efforts to utilize new energy sources, the United States, with its great 


supplies of coal and oil shale, has many alternatives. It has considerable geo- 
thermal energy potential, great competence in nuclear energy, and tremendous 
resources of forests and farms, and abundant usable solar energy. 

Most other countries are not so fortunate and they have limited choices. For 
example, a major result of the high prices and forthcoming depletion of oil has 
been to expedite a worldwide rush to go nuclear. Consider the position of a 
country such as France which has very limited deposits of coal and modest 
amounts of oil. It has rich soil, but this has been devoted to agriculture. If France 
wishes to enjoy amenities that energy can bring, it must expand nuclear facilities. 
This it has been doing, and it has developed local uranium isotope processes 
while pioneering in successful operation of a breeder reactor. 

Consider South America, which has little coal and which has oil of consequence 
in only a few countries. Much of the continent might use energy derived from 
plants and trees. But in practice, there is likely to be a tendency to follow the 
example of Brazil, which is going nuclear in a big way. Another country in the 
news is Iran, at the moment a large exporter of oil. However, the spectre of 
depletion is already apparent, and prudence seems to dictate to the Shah the 
necessity of acquiring nuclear plants and technology. Many others will follow. 
This is an unwelcome prospect for the United States. Can or will these pro- 
liferating installations be monitored effectively? Will terrorists acquire the 
making of nuclear weapons? In principle, the risks could be reduced by inter- 
national agreements involving inspections, and efforts to achieve safeguards 
should certainly be made. However, in practice the world is embarked on a 
dangerous course. 

For the long pull, a much more satisfactory approach applicable to many areas 
would be to harvest the sun's energy in a variety of ways. The amounts of solar 
energy impinging on the earth are tremendous. In even so profligate a country 
as the United States, annual use of energy amounts to about one thousandth that 
received each year from the sun. Some trees fix through photosynthesis as much 
as 2%-3% of the energy falling on them. Thus, in principle, United States 
energy needs could be met by devoting a small fraction of the land to tree farms. 
In the tropics, where growth conditions are often even more favorable, prospects 
for a large renewable energy source are excellent. If exploited, it would decrease 
pressure toward nuclear proliferation. It would appear desirable for the United 
States to encourage and assist such developments. 

One of the effects of the successful actions of the oil cartel has been to place 
very heavy burdens on the majority of the less developed countries — those 
that have little or no oil. The burden is at least twofold. They pay greatly 
increased prices for oil, but in addition their imports from countries such as 
Japan and those of Western Europe cost more. The example of the successful 
oil cartel plus financial strains has caused many countries to seek means of 
obtaining higher prices for their raw materials. Thus far such efforts have not 
been very successful. However, the lesson of oil has not been lost on companies 
in the United States. Some attention has always been devoted to assuring supplies 
of key raw materials. That attention is now keener than ever. Analysis show's 
that our principal vulnerabilities lie in the effects of sudden denial of some 
crucial material, such as chrome or tungsten. Adequate stockpiles of such ma- 
terials should be maintained. Given sufficient time, technology can overcome 
the effects of most shortages. One simple approach is to analyze the uses of a 
given substance in terms of the functions it serves. For some substances, it will 
be found that another, more abundant material can be used instead. 


Prospects for additional supplies of raw materials are mixed. In the mineral 
field, geologic knowledge is being advanced by the concept of plate tectonics 
and by the use of the Earth Resources Technology Satellite and deep sea 
exploration. Thus, discovery of new resources will be fostered. However, ability 
to exploit discoveries is handicapped by environmental constraints, by limited 
ability to exploit public lands, and by soaring costs of energy. The winning of 
mineral values is often energy-intensive. 

Today many of the materials employed are nonminerals — wood products, 
plastics, and other synthetics. The economy is especially dependent on items 
derived from natural gas and petroleum l The annual tonnage of such synthetics 
is somewhat less than that of steel, but in volume they surpass it. Most of the 
basic feedstocks for these materials could be derived from wood or plants. At 
one time the nation's supplies of ethyl alcohol, acetone, and butanol were pro- 
duced by fermentation of grain or molasses. The three chemicals are or could 
be used as a substantial fraction of the inputs for plastics and other synthetics. 
The key starting materials for fermentation are sugars. A convenient and com- 
paratively cheap one is glucose, which can be obtained from the starch of grains 
or the cellulose of trees by hydrolysis. Wood itself could be modified chemically 
to form a useful substitute for many plastics. These examples merely illustrate 
a part of the potential of renewable resources for replacing hydrocarbons. Demand 
for feedstocks currently derived from oil or natural gas could also be satisfied 
in considerable measure by chemicals from coal. Basic and applied research 
designed to exploit the feedstock potential of renewable resources and coal are 
likely to provide highly useful information and lead to important practical 
results. For several decades the existence of cheap petrochemicals served to make 
other sources noncompetitive and hence relatively uninteresting. The scarcity 
and high prices of hydrocarbons have created new opportunities. When the old 
sources are looked at in the light of modern technology by scientists employing 
present-day powerful experimental tools, new opportunities will become apparent. 

The vulnerability of the United States to stoppage of foreign imports could 
also be lessened if greater efforts were devoted to employing common materials 
such as those derivable from ordinary rocks. Such efforts would not only lessen 
the need to import, they would also be beneficial to the rest of the world. Thus 
the developments of the past two years are pointing toward efforts to use the 
abundant minerals of the earth's crust and materials that can be formed from 

Another major manifestation of exponential growth has been the population 
explosion. For a time, it seemed that the "Green Revolution" and other agricul- 
tural developments might keep pace with expanding populations. But even 
Mexico, the cradle of a successful Green Revolution, is now a net food importer. 
For a time it was an exporter, but the inexorable growth of population has eaten 
up the surplus. The situation varies greatly from region to region. Some of the 
less developed countries are still food exporters. Others are self-sufficient and 
possess untilled but suitable land. In most countries, yields are not nearly so 
high as in the United States. Thus it would appear that overall food production 
might be increased very substantially. However, the high cost of energy and 
fertilizers as well as social factors such as lack of economic incentives have 
restrained production. At the same time, population continues to grow. In conse- 
quence, world food production is no longer abundantly sufficient. Countries that 
were once major exporters no longer are. Today, North America and Australia 
are the only continents that are net exporters of food products. As for the 


United States, the situation has changed considerably during the past several 
decades. During the 'fifties and much of the 'sixties, there was an unwelcome 
surplus of food production leading to costly governmental programs to hold down 
acreage and to pay for storage of huge amounts of carry-over grains. During 
the 'seventies, matters changed drastically. Stockpiles of grains disappeared, idle 
land was farmed, and bumper crops were sold at substantial prices. The change 
in the demand for United States grain provides dramatic evidence that expanding 
world population is in the process of coming up against a barrier to further 
exponential growth — the food barrier. 

For the United States, this great turn of events is important. In the fiscal 
year ended June 30, 1975, this country exported agricultural products valued at 
$21.6 billion. This sum is comparable to the cost of oil imports in 1974 ($25.0 
billion), and indeed were it not for the huge export revenues obtained from 
agriculture the United States would have had problems with its balance of 

As we look ahead, a further tightening of world food supplies will increase 
revenues and provide an extremely valuable bargaining tool in the world political 
game. However, food shortages and perhaps famines are likely to put the United 
States in unenviable moral predicaments. At times this country may find itself 
having to decide which people will starve. Moreover, as United States imports 
of oil increase and as the oil cartel raises its prices, the United States might feel 
compelled to respond by raising prices of food exports to pay for the oil. This 
might give the United States the appearance of a nation that would squeeze 
funds from starving people in order to go for joyrides. 

Obviously, the United States must place first priority on putting its own 
house in order and caring for its own people. But the United States has enormous 
resources of land, raw materials, and trained people. Through improvements in 
agriculture, the sustainable yields could be substantially increased. Through 
improvements in forestry and forest management, very substantial increases — 
a factor of two or more beyond present production — might be attained. In addi- 
tion, both agriculture and forestry could ultimately enjoy additional factors of 
improvement by continuing basic research. Progress that is made in the United 
States is likely to be of worldwide application. Thus, we might assist others in 
the best fashion, that is, by helping them to help themselves. Another way we 
would do ourselves a favor while benefiting others is by using indigenous energy 
sources to such an extent that oil imports could be diminished. This would 
remove pressure from prices and extend the lifetime of world supplies. 

The problems of changing the mix of domestic energy supplies are largely 
economic, political, and environmental, with technological barriers of significant 
but moderate importance. In 1975 a large amount of basic and especially applied 
research was being carried out or financed by government, notably by the Energy 
Research and Development Administration and by the National Science Founda- 
tion. This work is important and should be pursued vigorously. However, the 
big bottleneck is the delay in applying on a large scale what is already known. 
For example, this country should implement well-known methods of conserva- 
tion. It should substitute coal for oil and gas. It should exploit resources of 
natural gas in tight geological formations in the Rocky Mountains and in the 
Appalachian Basin. It should proceed with construction of prototype facilities 
for synthetic fluids from coal. It should conduct full-size in situ retorting of oil 

In 1975, some people were beginning to understand that it would be prudent 
to take decisive steps with respect to energy, materials, and food. But there was 


by no means general awareness of the long lead times involved in effective 

That the United States and other countries face a decade and more of difficult 
adjustments to cessation of the kind of exponential growth characteristic of the 
'fifties and 'sixties is certain. But even the rough outline of the course those 
adjustments will follow is uncertain, for the crucial elements in determining the 
course are the people and their leaders. If the spirit of a people is crystallized 
by an event like the attack on Pearl Harbor, they will demand and get leader- 
ship. Then what previously looked like great problems become readily man- 
ageable. Before December 1941, the technological achievements of World War 

11 seemed impossible. Indeed, given this nation's condition in 1975, with clashing 
objectives and parochial demands, the accomplishments of World War II would 
require several decades. But such a situation cannot and must not prevail 

When the mood for decisive action crystallizes, the nation's leaders will again 
find it necessary to effect at least a partial mobilization of intellectual, scientific, 
technologic, and industrial resources. Obviously, the precise relationships will 
differ from those of World War II, for the problems and needs are different. But 
there is one lesson of World War II that should be remembered. The nation's 
scientists and engineers constitute an enormous intellectual reserve. During 
ordinary times, scientists pursue knowledge at its frontiers. Much of what they 
do looks esoteric, and indeed applications may be long in coming. But there is a 
highly practical side to most scientists. They are problem solvers. They are 
accustomed to analyze complex situations and to identify do-able pieces. Those 
that perform effectively are highly trained, experts in limited fields but possessed 
of broad fundamental knowledge. With their knowledge and experience, many 
scientists can rather quickly tackle new problems in fields quite different from 
their previous activities. 

These remarks about the adaptability of scientists are especially relevant to 
the Carnegie Institution. During World War II, contributions of the Carnegie 
staff such as Vannevar Bush and Merle Tuve to the national effort were enormous. 
In total, the contributions were of a magnitude far out of proportion to the 
numbers involved. This resulted in part because of excellence of the staff, but 
in addition it came about because of policies and traditions that have long 
determined the character of the staff and its work. The Institution has been 
guided by Andrew Carnegie's desire that the staff seek knowledge for the benefit 
of mankind. This guideline has found expression in the broad choices that 
individual staff members make. While giving top priority to seeking fundamental 
knowledge, they have maintained a broad awareness of societal needs and prac- 
tical applications. In consequence, over the years many important applications 
of Carnegie discoveries have been made. 

One way that scientists achieve and demonstrate excellence is by opening up or 
exploiting new areas for investigation. During the present era, progress in science 
has often been associated with opportunities created by new techniques, equip- 
ment, or instrumentation. Carnegie scientists, with their versatile and practical 
bent, have made more than their share of innovations in techniques, equipment, 
and in instrumentation. That contributions of this kind are continuing is 
demonstrated by the record of the past few years. This will be exemplified in the 
discussion that follows of highlights of the past year of work and events at 
the various departments of the Institution. 

The Year in Review 


In its mission of attempting to understand processes involved in the formation 
and evolution of the earth, the Geophysical Laboratory has often pioneered. 
Since its beginning in 1906, the laboratory has introduced many new concepts, 
techniques, and experimental equipment. As some questions were answered, new 
tools were created that led investigators to study yet other puzzles and indeed 
to formulate completely new topics for research. During the past few years, 
scientists at the laboratory have been particularly innovative with respect to 
equipment and technique. 

One example is the development this past year of equipment capable of 
achieving pressures of 500,000 atmospheres at a temperature of 2500°C. Such 
a pressure corresponds to that at about 1000 kilometers depth in the earth, 
whereas the temperature is greater than that of the white-hot filament of an 
electric light. Development of this equipment is a remarkable achievement 
in itself. However, what is perhaps more noteworthy is the creation of means of 
simultaneously measuring the pressure and temperature, and monitoring effects 
with spectrophotometers, x rays, and Mossbauer effect behavior of samples under 
these extreme conditions. These developments were conducted principally by Peter 
M. Bell and Ho-Kwang Mao. 

A striking feature of the pressure equipment is its low cost and simplicity. 
Most of the approaches to high pressure research have involved massive equip- 
ment with the experimental sample inaccessible to the observer while it is under 
high pressure. The new apparatus permits visual inspection and other direct 
observations of the sample while under pressure. The method of producing the 
pressure is to use the space between opposing diamond anvils. The sample is 
heated by absorbing the energy of a laser beam to which the diamond anvils 
are transparent. 

The new equipment makes the high temperature chemistry of the earth's deep 
interior accessible to study. Results from the new equipment will be followed 
avidly by theoretical geophysicists seeking to understand such processes of the 
earth's interior as those giving rise to the motion of tectonic plates. The new 
knowledge that will arise is also likely to lead to important practical results, 



much as the studies of P. W. Bridgman at Harvard were ultimately extended 
by the General Electric Company to the commercial synthesis of diamonds. 

Ho-Kwang Mao, David Virgo, and Carnegie Fellow Frank E. Huggins have 
designed, built, and tested Mossbauer equipment adapted to use in conjunction 
with the special conditions of the high-pressure diamond cell. Observations can 
be conducted successfully on 10-microgram samples. These scientists point out 
that ability to conduct Mossbauer studies at high temperatures and pressures is 
particularly important because of the ubiquitous occurrence of iron in the earth's 
crust and mantle. Moreover, iron is the only transition element of major 

Thus, iron can be expected to give rise to most, if not all of the complications 
associated with crystal-field and charge-transfer phenomena, variable valence 
and spin states, and the electrical and magnetic properties of cations in minerals 
likely to be found in the mantle. For this reason, Mossbauer studies of minerals 
containing iron under the pressures and temperatures of the mantle are expected 
to be especially informative. Furthermore, Mossbauer data as a function of 
pressure can be related to pressure-induced changes in the crystal chemistry and 
the magnetic, spin, and valence properties of iron in minerals, on which may be 
based interpretations of pressure-induced phenomena within the earth. 

Sometimes a substantial contribution is made by putting together in a new 
context simple techniques which when combined create a powerful new capability. 
This is true of a major advance in radioactive dating achieved at the Geophysical 
Laboratory with collaboration of the Department of Terrestrial Magnetism. 

Various clocks based on decaying radioactivity have been known for many 
years. Until about ten years ago their accuracy was about l%-2%. This was 
useful but, for example, in the dating of old rocks (3000 million years) an error 
of 30-60 million years was common. This level of accuracy is not adequate for 
many purposes. Subsequent improvement of electronic circuitry in mass spec- 
trometers at the Department of Terrestrial Magnetism and elsewhere created a 
potential for better measurement of isotope abundances. However, the potential 
was not being fully realized because of tiny amounts of adventitious impurities 
in reagents, in distilled water, and in the dust of laboratories. An example is the 
dating of zircons which contain uranium that decays to lead. In principle, the 
zircon can be dated by measuring the ratio of the atoms of lead that have been 
formed to the atoms of parent uranium now present. The amount of radioactive 
lead to be determined is often of the order of a few micrograms or less. The 
accidental addition of a few hundredths of a microgram of ordinary lead results 
in a serious error in the measurement. Quantities of this kind are ubiquitous, if 
only because of the large-scale use of tetraethyl lead in gasoline and consequent 
aerosols from motor exhausts. Thus laboratories, glassware, and reagents are 
contaminated. Through an effort sparked by Thomas E. Krogh with Gordon L. 
Davis and former Carnegie Fellow James M. Mattinson, techniques were devel- 
oped that cut the contamination to a low level. Reagents were purified by 
molecular distillation, and a new procedure using a Teflon bomb eliminated the 
need for extensive chemical processing. Preparation of a pure sample of 205 Pb to 
use as a quantitative tracer was also helpful. The net effect was that for 
the first time Krogh and colleagues could measure mineral ages of 3000 million 
years with an accuracy of ± 1-2 million years. The equipment involved is 
simple and cheap and will be duplicated elsewhere. It will facilitate accurate 
dating and detailed reconstruction of important Precambrian events, some of 
which hold the key to early evolution of life and some of which created many of 
the world's most important mineral deposits. 


For many years petrologists have been well aware of the role of comparatively 
small amounts of water in altering the properties of rocks, particularly at the 
high temperatures and pressures prevailing in the earth's interior. Among the 
effects are large changes in strength and melting points. In addition, water can 
cause changes in chemical relationships among the silicates. Water also probably 
has a role in explosive volcanism. In spite of this background of knowledge of 
one volatile, H 2 0, petrologists have not fixed much attention on a second 
volatile, C0 2 . 

When the role of volatiles is considered, emphasis is usually placed on water, 
not only because of the large amount in the oceans but also because of its 
abundance in some rocks. The working hypothesis is that the water now in the 
oceans originated deep in the earth. Because the amount of water in the oceans 
is much greater than the amount of C0 2 in carbonate rocks, attention has focused 
on water. However, if one takes seriously the concept of tectonic plates and the 
attendant underthrusting, one would be led to assume that large amounts of 
carbonate have been and are being recycled. Indeed, in areas where tectonic 
plates have collided, the amount of C0 2 underthrust as carbonates might exceed 
that of H 2 present in any form. 

During the past two years, David H. Eggler has placed a spotlight on the 
roles of C0 2 and has opened up what may well become a most significant area 
for research and understanding. Eggler has found that C0 2 dissolves in silicate 
melts and that it plays a significant part in the fractionation of magma toward 
silica-undersaturated compositions. The solubility of C0 2 is dependent on pres- 
sure. Thus a sudden release of pressure could give rise to a sudden release of 
C0 2 with important physical and chemical sequelae. 

In this year's Report, Hatten S. Yoder, Jr., Director of the Geophysical 
Laboratory, has outlined some results of this year's work on C0 2 . Carnegie 
Fellow Bj0rn 0. Mysen obtained infrared spectra on quenched C0 2 -saturated 
liquids and demonstrated that C0 2 enters these liquids both as discrete C0 2 
molecules and as the carbonate anion. Increasing temperature increases the 
proportion of total carbon entering the melt as carbonate anion, as does increas- 
ing basicity of the silicate melt. Highly polymerized (acid) melts such as the 
melt of albite composition dissolve the least C0 2 (1-2 weight percent) according 
to Mysen, and at temperatures near the albite-C0 2 liquidus the only carbon 
dioxide species in the melt is molecular C0 2 . Less polymerized silicate melts such 
as the melt of diopside composition display a large C0 2 ~ 3 absorption band under 
all conditions investigated. Infrared measurements indicate that the C0 2 ~ 3 ion 
is more important in hydrous than in H 2 0-free silicate melts. Increasing basicity, 
temperature, and pressure also result in increasing total C0 2 solubility in all 
compositions studied (CaMgSi 2 O , NaAlSi 3 8 , NaAlSi 2 6 , NaAlSiO.*, natural 
olivine nephelinite, tholeiite, and andesite) . 

On the basis of his present studies, Mysen places emphasis on the low solubility 
of C0 2 compared with that of H 2 and also on how strong the dependence of 
C0 2 solubility is on temperature, pressure, and silicate chemistry. These results 
together with phase equilibrium data on model mantle C0 2 -H 2 systems can 
aid in understanding the fractionation of tholeiitic magmas to compositions con- 
taining normative nepheline and larnite. It is proposed by Mysen that C0 2 -rich 
vapors can evolve from silicate magmas during ascent in the upper mantle. The 
development of such a C0 2 -rich vapor presents a means of generating nephelinitic 
magmas by partial melting of a peridotite upper mantle; previous models have 
lacked the mechanism to generate a C0 2 -rich vapor. 

The origin of kimberlite, containing initially the three principal minerals 


olivine, calcite, and phlogopite, depends greatly on the role of C0 2 and H 2 0. 
In the course of studying the system CaO-MgO-Si0 2 -C0 2 , Eggler discovered 
another invariant point that bears directly on the generation of kimberlite. The 
invariant point is formed by the intersection of the decarbonation reaction 
dolomite -f- enstatite = forsterite -f- diopside -\- C0 2 with the melting curve of 
the assemblage diopside + forsterite -\- carbonate at a pressure between 25 and 
30 kbar. Because of these relationships, the subsolidus phase assemblage of a 
simple peridotite composition at 30 kbar is not diopside' -f- forsterite -f- ortho- 
pyroxene -f- C0 2 but forsterite -j- orthopyroxene + diopside + dolomite. The 
latter assemblage melts to a carbonate-rich liquid at a temperature hundreds 
of degrees below the solidus of volatile-free peridotite. Pressure release by 
tectonic processes could, therefore, trigger melting that yields kimberlitic liquids 
at relatively low temperature. It had previously been thought that such melts 
could form only by fractionation of more silica-rich partial melts. Eggler also 
proposes, on the basis of these phase equilibria, a new concept of the low-velocity 
zone beneath the oceans where partial melting results from the presence of small 
amounts of C0 2 rather than H 2 0. 


George W. Wetherill became Director of the Department of Terrestrial Mag- 
netism in April 1975. He came from the University of California at Los Angeles 
where he had been Professor of Geophysics and Geology. Wetherill was a staff 
member at the department from 1953 to 1960 before going to UCLA. 

A graduate of the University of Chicago, Wetherill received bachelor's degrees 
in philosophy (1948) and physics (1949). He earned the M.S. degree in 1951 and 
the Ph.D. in 1953, both in physics. From 1951 to 1953 he was an Atomic Energy 
Commission Predoctoral Fellow, doing thesis research on spontaneous fission 
of uranium and thorium and nuclear processes in nature. Before that he was a 
research assistant at the University of Chicago, where he worked on the 170-inch 

While he was a staff member at the Department of Terrestrial Magnetism, 
Wetherill participated in a joint DTM-Geophysical Laboratory project to re- 
determine the decay constants of rubidium and potassium — standards that are 
fundamental to the measurement of geologic time. Wetherill's main contribution 
to this effort was his determination of the concentration and isotopic composition 
of inert gases such as argon by mass spectrometric isotope dilution. His formula- 
tion of the systematics of complex data for open systems is the basis of modern 
age determinations using the lead-uranium method. He has contributed exten- 
sively to the definition and assignment of age provinces in ancient Precambrian 
shield areas. 

Wetherill has been a principal investigator of lunar samples from Apollo 11 
to 17 since 1969. He has done major theoretical work on the origin of meteorites 
by studying the dynamics of stray bodies in the solar system. 

Ellis T. Bolton, the former Director, continues with the Institution as a senior 
staff member. He is currently Carnegie Institution of Washington Professor of 
Marine Studies at the University of Delaware. 

Bolton succeeded Merle A. Tuve as Director of the Department of Terrestrial 
Magnetism in 1966. He had joined the Carnegie Institution in 1949 as a fellow 
at DTM and had served as staff member, chairman of the Biophysics Section, 
and associate director. Although his training at Rutgers University (B.S., 1943; 
Ph.D., 1950) was in immunology and its use in indicating the degree of related- 


ness between species, at DTM he shifted his attention to biosynthesis in bacteria 
and in small molecules that are the precursors of proteins and nucleic acids. His 
work in these areas helped clarify the pathways by which sulfur is incorporated 
into amino acids and the role of the Kreb's cycle in the incorporation of carbon 
from C0 2 into amino acids and nucleotides. 

Bolton also contributed to development of the competition method for analysis 
of the inhibition of synthesis now called "feedback" control. Later he turned to 
studies of macromolecules and the more complex particles found in cells. He 
demonstrated the role of magnesium in stabilizing ribosomes, developed means of 
purifying ribosomes and animal virus, and showed the difference in composition 
between different sizes of ribosomes. An important contribution of Bolton's work 
was the development, with B. J. McCarthy and B. H. Hoyer, of an agar gel 
method for isolating single strands of DNA. This technique for analyzing rela- 
tionships between living things has created new possibilities for study in medicine, 
differentiation, taxonomy, and evolution. 

At the Mariculture Laboratory of the College of Marine Studies, University 
of Delaware, Bolton is part of an interdisciplinary group of biologists, chemists, 
and engineers engaged in research and development efforts to determine feasibility 
of large scale growing of filter- feeding marine invertebrates — oysters. The study 
is progressing well. 

Bolton commenced his work at Delaware in September 1974. From that time 
until April 1975, L. Thomas Aldrich served as acting director of the department. 
He carried on cheerfully and effectively, managing to perform vital administrative 
functions while maintaining his personal research program. 

This department has a long history and tradition of developing and main- 
taining excellent instrumentation. Much of the current research is made possible 
by its special facilities. 

For example, the mass spectrometers at the department have capabilities 
equaled by few elsewhere. Their ion sources have been so improved that high 
precision isotopic measurements can be conducted on a few nanograms of sample. 
The electronics of the instruments are excellent and are coupled with computer 
averaging techniques. In consequence, it is possible to measure isotope ratios to 
a standard deviation of .005 percent. This capability together with the techniques 
developed by Thomas E. Krogh and Gordon L. Davis at the Geophysical Labora- 
tory make feasible new accuracy in zircon dating. The capability also facilitated 
obtaining significant findings on the strontium isotope content of mid-ocean 
ridge basalts. 

According to plate tectonic theory, new oceanic lithosphere is produced at 
mid-ocean ridges by addition of material from deeper within the earth's mantle. 
The surface manifestations of this material are the mid-ocean ridge basalts 
(MORB), obtained by dredging and drilling of the sea floor. Previous work, 
primarily based on young dredge haul samples, showed that MORB is dis- 
tinguishable from other basalts, by both its trace element concentrations and 
its strontium isotopic composition. It was not clear from this earlier work that 
the difference was an essential characteristic of MORB or reflected biased 
sampling of rocks of a particular age. Stanley R. Hart reports results from 
older (15-40 million year) samples of MORB from the South Atlantic, obtained 
by drilling, which show that the characteristic differences between MORB and 
other basalts extend back in time. 

These differences between MORB and other basalts such as those produced 
in volcanic island arcs are examples of a fundamental problem which dominates 
current geochemical discussions of the mechanisms by which material is trans- 


ferred from the deeper interior of the earth to the surface regions: that various 
basaltic rocks, all presumably derived from the mantle, exhibit different chemical 
and isotopic characteristics. Further examples are given in the report by Hart 
and Carnegie Fellow William M. White, collaborating with J. G. Schilling, 
concerning basalts from the mid-Atlantic ridge, the Azores, and Iceland. The most 
obvious explanation is that the mantle is chemically heterogeneous on both a 
local and a global scale. If true, a new field of "mantle geology" directed toward 
an understanding of the basic principles governing these heterogeneities is opened. 

Further examples of the capabilities of the department staff in developing 
instrumentation are represented in contributions by I. Selwyn Sacks. Details have 
been described earlier. His wide amplitude, wide frequency seismometer and 
highly sensitive strainmeter are achieving increasing acceptance. The potentiali- 
ties of the strainmeter are still being investigated, but the indications are that 
it will have very important practical applications. This year ten of the meters 
were installed by the Japanese Government in and near a probable earthquake 
site. The Japanese are urgently seeking means of predicting earthquakes, and 
their assessment of the potential usefulness of the Sacks strainmeter in compari- 
son with other approaches has been favorable. 

The wide amplitude, wide frequency seismometers have made feasible an 
investigation conducted this year on seismic events attending the subduction 
of tectonic plates. 

The tectonic environment of the Andes and the adjacent portions of the 
South Atlantic Ocean are complementary to those of the mid-ocean ridges. 
Here down-going slabs of oceanic lithosphere plunge hundreds of kilometers 
beneath the continental lithosphere, resulting in the seismically active "Benioff 
zone," volcanism, and the formation of mountain belts. The relationships between 
these processes have been the subject of a continuing combined geophysical and 
geochemical investigation. Sacks and Carnegie Fellow J. Arthur Snoke report 
seismic studies of lithospheric plate motions in both South America and Japan. 
The seismic work suggests a new hypothesis for the generation of the thick 
lithosphere which distinguishes continents from ocean basins: that the cool 
down-going slab of oceanic lithosphere freezes out aesthenosphere in the "wedge" 
defined by the intersection of the continental plate and the down-going oceanic 
plate. In combination with migration of the ocean trench associated with the 
down-going slab, this could provide a mechanism for increasing the mass of 
continental lithosphere over geological time. 

Vera C. Rubin, W. Kent Ford, Jr., and Norbert Thonnard have drawn a very 
interesting conclusion from their study of the apparent anisotropy in the distri- 
bution of a particular class of galaxies as indicated by both optical and radio 
data: that the best interpretation of their observations requires our galaxy, as 
well as its local neighbors, to be moving at a velocity of about 500 kilometers 
per second with respect to the reference frame defined by the distant galaxies. 
The direction of this motion is perpendicular to the direction of the nearest large 
cluster of galaxies, the Virgo cluster, and the velocity corresponds to the orbital 
velocity of a body at the distance of our galaxy from the Virgo cluster. There- 
fore, the data are consistent with the idea that in some sense our galaxy is in 
orbit about the Virgo cluster, although not in the usual way, as the 10 11 year 
period of such an orbit is much greater than the age of the galaxies. These 
observations are inconsistent with the generally accepted isotropic distribution 
of 3°K cosmic blackbody radiation background. The discrepancy appears to be 
outside the observational errors of both sets of measurements, and represents 
an important problem for further investigation. 


Timely accomplishment of the observations mentioned above would not have 
been possible without the Carnegie image tube developed by Ford at DTM. 
Continued improvements in these devices and systems for utilizing them have 
been made by Ford. In addition, he reports advances in developing a new type of 
detector especially suitable for photometric work. This detector eliminates use 
of the photographic plate and provides the imaging data in digital form. 

In atomic physics, the intensity of production of characteristic x-ray line 
spectra by heavy ion bombardment yielded a simple exponential dependence on 
the atomic number of the target atom. Surprisingly, it appeared that current 
theories do not yield this result, even though its simple form strongly suggests 
that a theoretical explanation should be possible. Louis Brown and his co- 
workers have derived a simple expression which fits the observed exponential 
dependence of the line spectra on Z, as well as predicting the x-ray energy 
spectrum of the continuum radiation. Although this does not yet represent a 
complete theory for the problem, the agreement with observed data indicates 
that the approach taken is on the right track. 

In nuclear physics, the nuclear structure of 16 and 10 B were successfully 
investigated by formation of the compound nucleus by bombardment of 15 N 
and 9 Be with polarized protons. The former nucleus is of astrophysical interest as 
15 N(p,a) 12 C and 15 N(p,y) 16 are competing steps in the CNO cycle of stellar 
energy production. 

During this report year, organized activities of the Biophysics Section were 
concluding. This meant the ending of a highly successful, highly productive 30 
year experiment. This year's Report of the Department of Terrestrial Magnetism 
contains a compact history of the experiment written by Richard B. Roberts. 
In it he outlines the origins of the Biophysics Section, describes its evolution, and 
points to some of its major accomplishments. 

At the conclusion of World War II, Tuve became director of the department 
and brought about a reorientation of its program. In particular, nuclear physics 
had received a tremendous emphasis from the development of the atom bomb 
and seemed destined to be carried out by large groups with unlimited funds. In 
contrast, biophysics appeared to be more suitable for small groups and (with the 
advent of tracers) due for a burst of progress. In addition, prewar experience 
had demonstrated that physicists could contribute to biology not only in tech- 
nique but in the planning of experiments. Tuve therefore decided to reduce the 
emphasis on nuclear physics and to initiate a Biophysics Section. At this initial 
stage the group included Abelson (chairman), Dean Cowie, and Roberts, all 
trained in physics. In addition, the group was fortunate to have with them H. H. 
Darby and W. R. Duryee as guest investigators who provided a much needed 
wealth of biological experience. 

The experiments of the first years, 1947 to 1951, were largely exploratory; 
they included such diverse topics as capillary permeability, and the effects of 
ultraviolet radiation, vitamin B 12 , and various ions on biological material. The 
biological materials used ranged from humans to bacterial viruses, but following 
Robert's participation in the bacteriophage course at Cold Spring Harbor there 
was a steadily increasing use of the bacterium Escherichia coli as biological 

By 1952 the group had been expanded with the addition of Bolton and Roy J. 
Britten and interest of the group had focused on the synthesis of amino acids and 
nucleotides by E. coli. This subject was studied intensively through 1955, cul- 
minating in the publication of the book Studies of Biosynthesis in E. coli. This 
book was well received generally and was considered to be a "bible" by some 


microbiologists. From that time on, the Biophysics group was an important 
contributor to the flourishing field of molecular biology during its most exciting 
years. In his write-up, Roberts points to the productivity of the small Biophysics 
Section. One measure is three books and 200 or more papers. 

A better measure lies in the listing of significant findings: (1) iso topic compe- 
tition method; (2) flow patterns in metabolism in E. coli; (3) feedback inhibition; 
(4) demonstration that nascent protein is attached to ribosomes; (5) virus 
purification on DEAE columns; (6) patterns of ribosome synthesis; (7) proof 
that newly formed RNA in bacteria is one-third messenger, two-thirds ribo- 
somal precursor; (8) agar column for nucleic acid reactions; (9) messenger isola- 
tion; (10) application of nucleic acid interaction to measure relatedness of species 
and degree of divergence; (11) homology of DNA in E. coli and lysogenic virus; 
(12) detailed studies of divergence in virus, bacteria, algae, vertebrates, rodents, 
and primates; (13) application of nucleic acid reactions to cancer virus investiga- 
tions; (14) demonstration that DNA remains identical but RNAs differ in 
different tissues and stages of development; (15) extensive studies of nucleic 
acid reaction kinetics; (16) discovery of repeated DNA; (17) properties of 
repeated DNA; (18) involvement of catecholamines and peptides in learning 
and memory. 

At the conclusion of his summary, Roberts ends on this thoughtful note. 

"We are pleased to have participated in this exciting period in the development 
of biology. We believe that we did make significant contributions and that, since 
some of us will carry on in different places, our history is not complete. We are 
especially pleased to note the contributions being made by 22 Fellows who 
received a part of their training with us." 


It is premature to celebrate the successful completion of the Irenee du Pont 
Telescope, for there is work that remains. However, most of the tasks have been 
completed and very successfully. For example, the figuring of the main mirror 
has been finished, and tests show that it is excellent. The mirror has been shipped 
to Chile and it has arrived on Las Campanas safely. The telescope itself was 
thoroughly shop-tested in Los Angeles, subsequently shipped to Chile, and in- 
stalled without problems in its dome on Las Campanas. 

Construction of a major observatory on a mountain 5000 miles distant is a 
demanding task that requires close coordination of many engineering and con- 
struction activities. The telescope was designed by the staff of the Hale Observa- 
tories, led by Bruce H. Rule as chief engineer and project officer. The late Ira 
S. Bowen made innovative contributions to the optical design. Testing of the 
telescope and design of auxiliary instruments are, as expected, requiring much 
additional effort on the part of staff members. Large demands are also made 
on the administrative staff. 

A novel control system for the du Pont Telescope has been created by J. T. 
Fridenberg of Caltech's Astroelectronics Laboratory with close collaboration of 
staff members of the Hale Observatories. A prototype has been tested on the 1.5 
meter telescope on Palomar Mountain. 

The telescope automatic control system was designed to control large observa- 
tory telescopes with a degree of sophistication equal to that provided by a typical 
general purpose computer. By making use of modern microprocessor modules, the 
control system operates the telescope independently of the data acquisition or 


other associated functions. This distributive processing technique isolates the 
control functions from problems often encountered with single computer control. 
A telescope so controlled avoids operational interruptions that normally occur 
when a computer control system is off-line. 

In addition to responsibilities in connection with the new Irenee du Pont 
Telescope and to meeting requirements for new equipment, the staff must devote 
part of their time to ensuring that the various existing telescopes and their many 
pieces of auxiliary apparatus are maintained in good condition. This is essential 
if both the local astronomers and the many guests are to function effectively. In 
spite of these burdens, which are shared by a staff of limited size, the research 
productivity of the Hale Observatories is excellent. This year much attention has 
been devoted to studies of galaxies — their structure, dynamics, clustering, and 
evolution. Horace W. Babcock, Director of the Hale Observatories, notes that a 
review of research topics reported upon for the past year shows that some twenty 
different investigations related to galaxies are in progress or have just been 

A galaxy such as our own, which includes the sun and the Milky Way, con- 
tains many millions of stars that occupy a relatively flat, disk-shaped region of 
space. Edwin P. Hubble classified the shapes of galaxies according to a bifurcated 
scheme with a sequence of forms branching from elliptical into normal spirals 
and barred spirals. Later Olin J. Eggen, Donald Lynden-Bell, and Allan R. 
Sandage introduced the concept that galaxies were formed by contraction and 
collapse of a primordial gas cloud to a plane. 

J. Richard Gott III and Trinh X. Thuan of the California Institute of Tech- 
nology have developed a more detailed analysis of events during galaxy formation. 
They analyzed the competing effects of dissipation and turbulent viscosity, using 
a model of cloud formation in the early protogalaxy. Numerical models, including 
both stars and gas, were used to simulate the collapse. The stars and gas collapsed 
together until they reached the plane, whereupon the gas formed a disk and the 
stars continued in elliptical orbits to form a spheroidal halo. Some oscillations 
of the disk are expected before equilibrium is reached. Models may be used to 
relate galaxial parameters to the original size and angular momentum of the 
cloud. In general, if star formation in the early protogalaxy is essentially com- 
plete by the time of maximum collapse, an elliptical galaxy is formed. If a 
significant amount of gas remains at this stage, it will dissipate its energy to 
form a disk-like spiral. Observed properties of ellipticals and spirals imply 
that ellipticals formed out of larger perturbations of the primordial gas cloud 
than did spirals, giving the ellipticals greater densities and shorter collapse 
times. If early star formation is rapid and dependent upon the square of the 
density, ellipticals are expected to complete their star formation by the time of 
maximum collapse, while spirals are not. Protogalaxies arising from larger initial 
density fluctuations cluster more readily, accounting for the preferential occur- 
rence of ellipticals in large clusters. 

The best hope for advancing knowledge of the chemical composition and 
chemical evolution of galaxies lies in the interpretation of the spectra of the 
regions of ionized interstellar gas within them. Leonard T. Searle has recently 
surveyed the strength of emission lines that are probes of ionization structure 
in a number of H II regions in the galaxies M31, M101, and M51. He has carried 
out a systematic survey of the behavior of the infrared lines of [S III] and is 
collaborating with G. Shields of the University of Texas in the interpretation 
of the data. The observed behavior of the [S III] lines is not in accord with the 


predictions of the simple ionization-structure models that have so far been used 
to interpret the spectra of H II regions. Unexpectedly, large density fluctuations 
within the emitting regions are indicated. It is already clear that the low excita- 
tion that is characteristic of regions situated in the inner spiral arms of galaxies 
is not simply a consequence of an enhanced oxygen abundance, as had earlier 
seemed likely. 

Observations bearing on the distribution of metallicity in the Andromeda 
Spiral, M31, were obtained by M. Schwarzschild of Princeton University in 
collaboration with J. Beverley Oke. They used the multichannel spectrometer on 
the Hale Telescope to obtain energy distributions at several locations in the 
central bulge of M31. The results can be interpreted as an increase of the heavy 
element abundance toward the nucleus of the spiral. Beyond a radius of 2', no 
further abundance changes were observed; but at 7'.5 and 10' from the center 
of M31 along the major axis, there is evidence for young hot stars. 

Sandage completed analysis of redshifts and photometry of selected groups 
and clusters of galaxies in the southern hemisphere. He found that when aggre- 
gates of galaxies having populations in the range of 20 to 300 are examined, 
the mean absolute magnitude of the first-ranked galaxy varies by only about 
0.15 mag. This suggests that the absolute magnitude of the first-ranked galaxy 
is determined by some physical upper limit in the galaxy formation process that 
is independent of the statistics relating to other members of the cluster. 

James E. Gunn and Oke have continued their separate programs to discover 
faint clusters of galaxies and to measure the redshifts and absolute spectral- 
energy distributions of the brightest members of these clusters. The first aim 
is to produce a list of clusters found under well-controlled conditions within a 
definite limit of apparent magnitude. Hundreds of clusters were identified on 
plates obtained in the course of an initial survey with the 1.2-meter Schmidt 
telescope. Spectrophotometric observations of a few show that redshifts are 
typically between 0.20 and 0.35. A deeper survey, conducted with a 90-mm image 
tube at the prime focus of the 5-meter telescope, has produced a sample of about 
30 very faint clusters of galaxies that typically have redshifts larger than 0.40. 

In addition to work on galaxies, other studies were conducted at the Hale 
Observatories. A few examples follow. 

George W. Preston, Arthur H. Vaughan, Jr., and former Carnegie Fellow 
Richard E. White have completed their investigation of Hg isotopy anomalies 
in the mercury stars. Wavelengths of Hg II A 3984 in 30 stars are distributed 
rather uniformly from the value for the terrestrial mix of isotopes to a value that 
implies a preponderance of 204 Hg. The wavelengths are loosely correlated with 
effective temperature. Relative isotopic abundances derived from partially re- 
solved profiles of A 3984 in i Coronae Borealis, x Lupi, and HR 4072 suggest that 
simple mass-dependent fractionation has occurred in all three stars. This indicates 
that an isotope separation process rather than exotic nucleosynthesis is responsible 
for the isotope anomalies. Accordingly, a scheme has been devised whereby 
isotopic compositions can be inferred from a comparison of stellar wavelengths 
and equivalent widths of A 3984 with those calculated for a family of fractionated 
isotopic mixes. For one star, 46 Aquilae, Preston reports a wavelength for the 
Hg II line that corresponds to pure 204 Hg. 

Jesse L. Greenstein has continued the colorimetric search for white dwarfs, 
especially those with late-type spectra and colors, using the multichannel spec- 
trophotometer on the Hale Telescope. The rarity of yellow and red degenerate 
stars continues to contradict the simple theory of cooling of the core of a 


degenerate object. Although numerous objects have been found cooler than 
8000°K, none have been found cooler than about 4500°K (EG 202). A few have 
relatively strong metallic lines, but at wavelengths longer than A 5500 they 
prove to be relatively blue objects. The reasonable possibilities seem to be that 
(1) the cooling becomes rapid below effective temperature 5500°K and the stars 
disappear into the black dwarf stage, or (2) some physical mechanism delays 
cooling of old white dwarfs. 

James A. Westphal and Jerome Kristian have continued to refine and test 
the application of SIT television camera tubes for two-dimensional (image) 
photometry. The level of photometric precision obtained at the telescope is close 
to the 1% limit expected from laboratory tests of the system. Over a range 
of 10 magnitudes, from 12 to 22, it is possible with the SIT-Vidicon to reproduce 
with highly satisfactory precision a broad-band sequence established with 
photomultipliers. The limits are not intrinsic, but at magnitude 22 the signal 
is completely dominated by the sky. SIT measurements should be accurate to 
fainter levels, although fainter photoelectric standards are not available. In a 
ten-minute exposure, stars are measurable fainter than V = 24 mag, with 
progressively lower precision attributable to photon statistics. Currently, the 
1% limit for precision of the SIT system is due primarily to seemingly random 
variations of the order of 0.5% in the background. 

The same SIT system used in image photometry has been employed by 
Westphal and Kristian for several programs of low- and high-resolution spectros- 
copy. When used with a small spectrograph, there are two important features 
of the SIT detector that make it particularly suitable for faint objects. The 
reproducibility of the transfer function allows accurate sky subtraction, while 
the two-dimensional format permits the spectra of the sky and object-plus-sky 
to be recorded simultaneously. The sky-subtraction process functions well and 
produces difference spectra that are limited essentially by photon noise. The 
SIT-Vidicon has also been applied successfully as a detector at the coude 
spectrograph of the Mount Wilson 2.5-meter telescope, where it has been used 
by Robert J. Brucato to study stellar line profiles and by C. P. Wilson to 
measure velocity dispersions and rotations of the central regions of galaxies. 
Cooling of the tube to — 90°C makes possible exposure times of several hours. 

An experimental grating spectrograph equipped with a cooled integrating SIT- 
Vidicon camera is nearing completion under Gunn's direction. The system is 
being interfaced with a PDP- 11/40 minicomputer to give real time capability 
for addition, averaging, reduction, and display of spectra. 

A new f/60 infrared photometer has been put into operation on the Mount 
Wilson 1.5-meter telescope. Signal modulation is obtained by wobbling the 
Cassegrain secondary of the telescope. The system provides a very satisfactory 
level of performance for the study of infrared radiation from molecular clouds, 
H II regions, and external galaxies. 


Located not far from the center of the Stanford University campus, this 
department is ideally situated for botanical research. Within easy range are 
such diverse environments as the hot, dry Death Valley; the cool, moist coastal 
zone; and the alpine Sierras. The department has transplant gardens in these 
regions and an extremely well-equipped mobile laboratory to utilize the diverse 
plant resources. 


On the Stanford campus, the department has extensive facilities on a 10-acre 
plot of ground. This year comprehensive renovation of the headquarters building 
and construction of new facilities were completed, providing excellent additional 
laboratory space. The department is now in position to interact even more effec- 
tively with Stanford and is already doing so through joint efforts on teaching 
and research. 

The total joint enterprise has at its disposal unique experimental capabilities. 
These are in part a result of collaboration between Olle Bjorkman and Joseph 
Berry of the department and Professor H. A. Mooney of Stanford. The equipment 
includes controlled growth chambers and a mobile laboratory. Thus it is possible 
to duplicate at the laboratory most of the various natural conditions that plants 
encounter, ranging from those of Death Valley to the conditions of the coastal 
regions. In addition, it is possible to create combinations of environmental 
variables which do not exist in natural environments, such as extremely high 
temperatures together with very high humidities or high C0 2 concentrations. 

The mobile laboratory has all the measuring and control systems needed for 
continuous and simultaneous measurement of such parameters as rates of C0 2 , 
2 , and H 2 vapor exchange and a host of other relevant variables, and is also 
equipped with a small dedicated computer. A most valuable feature of the on- 
board data acquisition and processing system in addition to providing greatly 
improved speed, flexibility, convenience, and accuracy in data collecting is that 
the immediate processing of the data enables observation of important plant 
responses to a change in a certain factor or combination of factors with sufficient 
speed to permit further study and manipulation of these responses. 

When not in use in the field, the mobile laboratory is docked next to the 
controlled growth facilities and is there employed as an integral part of the 
stationary laboratory. Having a single sophisticated instrument package which 
can be used both in laboratory and in field eliminates extensive cross-calibration 
efforts, saving both time and funds, and giving the same precise base line to 
measurements made under both circumstances. 

These major items are only part of the resources at the Department of Plant 
Biology. The completion of new construction has provided additional laboratory 
space. At the same time, David C. Fork and Carnegie Fellow Tetsuo Hiyama 
have developed and built new pieces of equipment that permit detailed examina- 
tion of processes of photosynthesis with control of intensity, wavelength, and 
time to periods as short as 2 X 10~ 7 seconds. This equipment, together with 
apparatus developed earlier by C. Stacy French, represents an inventory of 
powerful tools for investigation of photobiology. 

During the report year, department staff lived with the disruption attendant 
to extensive construction and renovation. Nevertheless, the staff enjoyed a 
productive year in teaching and research. 

Bjorkman, together with Mooney and James Ehleringer, a Stanford graduate 
student, continued studies in progress last year on the photosynthetic properties 
of plants from contrasting thermal environments. Direct measurement of photo- 
synthesis corresponded well with previous growth studies, indicating that the 
temperature dependence of dry matter production is closely related to the 
temperature dependence of photosynthetic capacity, whether the plant studied 
used the C 3 or C 4 pathway of carbon fixation, and whether it was a native of a 
cool coastal climate or the extremes of temperature found in the bottom of 
Death Valley. When the various plants were all grown under similar conditions, 
whether warm or cool, some showed the capability of adapting the photosynthetic 


machinery to growth under temperature regimes far removed from normal, while 
others did not. Plants which could perform excellently under the most extreme 
conditions, however, seemed the least adaptive. 

Further studies by Bjorkman indicate that the high temperature inhibition of 
photosynthesis is not caused by a decrease in stomatal conductance, nor by 
temperature-induced alteration of diffusive transport of C0 2 fixation sites within 
the cells. Neither is it caused by a general breakdown of cellular integrity, since 
inhibition of respiration cannot be detected until temperatures are raised signifi- 
cantly above those at which inhibition of photosynthesis is found. Under rate- 
limiting light intensities, however, the quantum yield for photosynthesis falls 
off sharply at the temperature at which inhibition of photosynthesis is first 
found. Likewise, the quantum yield of the extraordinary desert plant Tidestromia 
oblongifolia, which can flourish at a temperature of 47°C, is sharply reduced 
when the plant is grown at lower temperature, e.g., 22°C. The cause cannot be 
alteration of any barriers to C0 2 movement into the cell, nor reduction in 
chlorophyll, protein, or carboxylation enzymes. Bjorkman hypothesizes that the 
membrane properties which permit photosynthesis at high temperatures preclude 
their effective function at low temperatures. 

Fork, Joseph A. Berry, and Ms. Garrison then heat-treated leaves of Tidestro- 
mia and examined various properties of chloroplasts isolated from them. It 
was clear that photosystem II electron transport was strongly inhibited by 
treatments which did not affect photosystem I. The temperature at which this 
effect is first noted corresponds fairly well with the temperature at which 
inhibition of photosynthesis is first observed. Changes in the fluorescence kinetics 
of intact leaves also occur at temperatures at which thermal inhibition of photo- 
synthesis is found. These results also implicate system II. Overall fluorescence 
yield is also affected by temperature, but transitions in fluorescence yield occur 
at temperatures significantly above those at which damage to photosynthesis 
occurs. Likewise, studies on ion leakage show that cell permeability is also 
normal at temperatures above which photosynthetic damage occurs. It is clear 
that photosynthetic membranes are more sensitive than the plasma membrane 
to thermal injury. 

Carnegie Fellows Norio Murata and John H. Troughton, together with Fork, 
grew the blue-green alga Anacystis nidulans under different temperatures selected 
to bring about known variation in the fatty acid composition of their membranes. 
Studies using electron spin resonance spectrometry on isolated fragments suggested 
that in cells grown at higher temperatures, the marked change in the environ- 
ment of the spin label with increasing temperature occurred at higher tempera- 
tures. Maximum fluorescence yield also occurred at higher temperatures for cells 
grown at higher temperatures, as was the case for lamellar fragments or for 
structures formed by reaggregation of extracted galactolipids and chlorophyll a. 
Fluorescence of the extracted chlorophyll alone, however, or extracted phycobilins, 
showed no effect of growth temperature. Thus the changes observed in fluores- 
cence yield in intact cells must be produced by interaction between the pigments 
and membrane lipids. The changes are probably caused by transition of mem- 
brane lipids between the liquid-crystalline state and the mixed solid, liquid- 
crystalline state. 

Further studies on electron transport as indicated by P700 reduction, oxygen 
evolution, and the shift of the thylakoid membranes from pigment state 1 to 
pigment state 2 also gave clear evidence of the membrane phase transition at 
roughly the same temperatures. These studies complement those by the physio- 


logical ecology group on temperature effects on photosynthesis under conditions 
of thermal inhibition. 

Fork and Jeanette Brown have continued studies on light-induced shifts in 
carotenoid absorption in a wide variety of different algal groups. Included were 
representatives of a newly identified algal group, the Eustigamatophyceae, plants 
containing only chlorophyll a, /^-carotene, and a major xanthophyll. Both light- 
minus-dark difference spectra and kinetics of the changes were examined. It is 
clear that such shifts occur in virtually all photosynthetic organisms, with the 
blue-green algae standing as the single exception. Further study is needed, 
however, to understand the physicochemical events which underlie these changes. 

Brown has continued studies on the properties of chlorophyll-protein com- 
plexes isolated from higher plants. One of these, CPI, is rich in P700 and 
contains only chlorophyll a. CPI is considered to be the reaction center chloro- 
phyll. The other complex, CPU, contains all of the chlorophyll b, plus an equi- 
molar concentration of chlorophyll a. Studies with Suzanne Acker and Jacques 
Duranton at the French Atomic Energy Center, Saclay, near Paris, show clearly 
that CPI is metabolically isolated from the rest of the chlorophyll-containing 
membranes in the plastid. Its pigments are both synthesized and degraded more 
slowly than the rest of the pigments. 

Fluorescence emission spectra at room temperature and at — 190°C were 
measured for both CPI and CPU. Dilution of CPU did not alter the fluorescence 
emission spectra whether excitation occurred with 420 nm or 470 nm light 
(preferential chlorophyll a or chlorophyll b excitation, respectively). Thus the 
two chlorophylls are clearly on the same protein (though not necessarily on the 
same polypeptide chain), since otherwise dilution would have decreased energy 
transfer between chlorophyll b and chlorophyll a. 

William F. Thompson, who joined the department this year, has initiated 
extensive studies of plant nucleic acids, concentrating particularly on DNA 
hybridization techniques. Studies on reaggregation of pea DNA suggest that the 
genome is organized with many clusters of repetitive sequences. 

Winslow R. Briggs, Director of the Department of Plant Biology, has been 
able to resume work started last year in the laboratory of Professor Rainer 
Hertel, University of Freiburg, Germany, on a promising candidate for the 
phototropic photoreceptor. With a Visiting Investigator, Jack Freeberg, he has 
obtained evidence for a light-induced absorbance change in a membrane fraction 
from corn coleoptiles. The change, which decays in the dark, suggests the light- 
induced reduction of a flavoprotein with subsequent reduction of a 6-type 
cytochrome. Though other workers have seen such changes and related them to 
the phototropic photoreceptor, the present work represents the first demonstration 
of such changes outside the fungi, and the first suggestion that the pigment 
is membrane-associated. Robert Brain, a Stanford undergraduate visiting the 
laboratory during the summer of 1975, has been able to show the same changes 
in a membrane fraction from the fungus Neurospora crassa. 


In his introduction to the Report of the Department of Embryology, the 
Director, James D. Ebert, stresses the communal nature of the scientific enter- 
prise, emphasizing both the interactions of the department with other labora- 
tories and the interactions of staff members, fellows, and students within the 
department. He takes as one of his prime examples the way in which new 


insights into gene structure have emerged through the use of restriction enzymes. 
The accomplishments of the department in the isolation of genes, in which 
Donald D. Brown has pioneered, are well known. In significant advances re- 
ported this year, the newly emerging technology of restriction enzymes has played 
a major part. This technology was not developed in the department, but several 
staff members and fellows have been quick to apply it critically to their problems. 
In this continuing study of the "anatomy" of the DNA which codes for 28S 
and 18S ribosomal RNA (termed ribosomal DNA or rDNA), a pivotal role has 
been played by Peter K. Wellauer, a Fellow of the National Cystic Fibrosis 
Research Foundation, along with staff members Donald Brown, Igor B. Dawid, 
and Ronald H. Reeder. In Xenopus laevis, about 450 copies of these genes are 
present on each of two chromosomes in every cell. Earlier studies revealed that 
these genes can be isolated completely free of the rest of the cell's DNA. Studies 
in the department and elsewhere have shown that rDNA is composed of multiple 
repeating units which are tandemly linked end to end. About two-thirds of each 
repeating unit is a gene region which serves as the template for the transcription 
of 28S and 18S ribosomal RNA. The remaining one-third of each repeating unit 
is not transcribed into RNA, has no known function, and is called the "spacer" 

Within a cell, all 900 copies of the transcribed gene region appear to be 
identical, both in length and in nucleotide sequence, suggesting that cells have 
some mechanism for maintaining these hundreds of copies identical within close 
tolerances. The recent unexpected finding was that the spacer regions between 
the genes come in a variety of lengths. As many as seven different spacer sizes 
have been found on a single chromosome alternating with gene regions of 
constant length. 

The existence of heterogeneous spacer lengths in rDNA was discovered by 
the use of a site-specific restriction endonuclease called EcoRI. EcoRI is one 
of a newly described class of enzymes which recognize specific short nucleotide 
sequences in the DNA and will cut the DNA only at the site of those sequences. 
In rDNA, the EcoRI recognition site occurs just twice in each repeating unit. 
Cutting the DNA at these two sites produces two fragments per repeating unit — 
one fragment containing only gene sequences and the other containing all of the 
spacer plus some gene region on each end. Electrophoresis of these fragments on 
agarose gels separates them according to size and reveals that the spacer- 
containing fragments come in different size classes while the gene fragment is 
homogeneous in size. 

When EcoRI is used to cut rDNA isolated from individual frogs, it is found 
that the pattern of spacer length can vary from one individual to the next. This 
has made it possible to use rDNA spacer pattern as a genetic marker, that is, to 
select two frogs which have different spacer patterns, mate them, and observe 
how the spacer patterns are transmitted to the progeny. 

Dana Carroll, a Fellow of the U.S. Public Health Service, also working 
with Donald Brown, has used a similar approach in further investigating the 
nature and organization of heterogeneity in the oocyte-type 5S DNA of Xenopus 
laevis. Again, not only is the existence of heterogeneity confirmed, but it is 
shown to be located in the AT-rich portion of the spacer. 

In these examples at least, length heterogeneity results from the presence 
of variable amounts of reiterated simple sequences in nontranscriber spacer, 
with the region of variable length being close, if not adjacent, to the 3' end of 
the transcription unit. 


Carroll's findings are a part, albeit a key part, of a larger undertaking by 
Donald Brown and his colleagues, namely the exploration of the "dual 5S DNA 
system" of Xenopus. The Xenopus genome contains at least two kinds of 5S 
DNA. "Somatic" 5S DNA appears to be made in all cells. "Oocyte" 5S DNA 
synthesis is "turned off" in somatic cells. Key questions concerning gene regula- 
tions are thus posed. Donald Brown and E. Jordan describe further progress 
in isolating 5S DNAs; and in a study related closely to Carroll's, Carnegie 
Fellow Ronald Brown has analyzed in detail the spacer sequence adjacent to the 
3' end of the 5S DNA transcription unit. It is thought that the end of the gene 
itself and the first two spacer nucleotides may be at least part of the "termina- 
tion signal" for RNA polymerase. 

Significant progress is again reported by Douglas M. Fambrough and his 
colleagues, especially graduate student Peter Devreotes, U.S. Public Health 
Service Fellow Katherine Tepperman, and Professor Daniel Drachman of The 
Johns Hopkins University School of Medicine. 

For several years Fambrough and his co-workers have been focusing on the 
differentiation of the skeletal muscle plasma membrane and the organization 
of skeletal muscle fibers. Many of their studies have centered upon the 
acetylcholine receptors of skeletal muscle, since they appear in the plasma 
membrane as readily detected functional molecules early in muscle development; 
are later confined to a small area of cell surface, the postsynaptic surface of 
the neuromuscular junction; and in denervated skeletal muscle reappear all over 
the surface of each muscle fiber. Because acetylcholine receptors can be readily 
identified not only by their function but also through the use of a radioactive 
probe which binds to them with extreme tenacity (iodinated a-bungarotoxin), 
they can be studied in ways which are unique. The group's working hypothesis 
includes the following elements. Acetylcholine receptors are synthesized by the 
normal protein synthesizing machinery. Newly synthesized receptor components 
occur as organized macromolecular structures of molecular weight about 250,000 
daltons incorporated into internal cell membranes. These "precursors" are then 
inserted into the plasma membrane, the population of precursors being enough 
to support two hours of new receptor incorporation into plasma membrane 
without any new protein synthesis. Once in the plasma membrane, the receptors 
are functional molecules and are free to interact with the radiolabeled probe. 
Plasma membrane receptors are degraded by a random-hit process which is 
energy-dependent and proteolytic, probably involving internalization of mem- 
brane containing receptors and then degradation by lysosomal enzymes. During 
the past year, the group has developed techniques which will allow direct testing 
of several aspects of this model of the life history of receptor molecules. They 
have also devised a method for continuous monitoring of receptor degradation 
in organ-cultured skeletal muscle and have begun extending findings to adult 
mouse muscle. 

Drachman and Fambrough earlier investigated the disease myasthenia gravis. 
They have now studied another human disease, congenital myotonic dystrophy, 
in which the muscle pathology suggested the possible participation of denervation. 
Since the appearance of extrajunctional acetylcholine receptors is a sensitive 
marker for denervated skeletal muscle surface, Drachman and Fambrough have 
used the radiolabeled probe for acetylcholine receptors on biopsy specimens from 
patients with congenital myotonic dystrophy. Biopsy material from two patients 
with the denervating disease amyotrophic lateral sclerosis has also been examined. 

During 1974-1975, Richard Pagano was joined by two postdoctoral Carnegie 


Fellows, Leaf Huang and Masatoshi Takeichi. Together with Ebert's colleague, 
Keiko Ozato, joint Carnegie-Marine Biological Laboratory Fellow, they have 
vigorously extended their explorations of phospholipid-cholesterol dynamics in 
mammalian cells. Their objective is to understand the dynamic movements which 
cellular phospholipids and cholesterol can undergo in cell membranes and the 
effects which perturbations in these motions can have on cellular activity. Among 
their studies described in the body of the Director's Report, two are especially 
noteworthy. They have combined their earlier knowledge of the interactions of 
cell membranes and lipid vesicles with the technique of high resolution electron 
microscope autoradiography to determine the surface distribution of one lipid 
component, phosphatidyl choline, on Chinese hamster fibroblasts. Second, in 
cooperation with Ozato and Ebert, they have begun to examine the effects of 
cell surface lipid perturbations on lymphocyte activation. 

Pagano's and Huang's findings bear directly on the possibility of using lipid 
vesicles, into which appropriate macromolecules have been reconstituted, to 
"transplant" a foreign membrane protein (or other marker) into the plasma 
membrane of an intact host cell. They have been able to clarify the relative 
contributions of the various "pathways" a lipid vesicle can take in its interaction 
with a mammalian cell. The major pathways of lipid incorporation in vesicle- 
treated cells are fusion of vesicle and cell and exchange of lipids between vesicle 
and cell. Pagano and Huang have established that the fusion process is dominant 
at 37°C, while at 2°C (or with energy depleted cells), exchange of lipids between 
vesicles and cells is important. Finally, they have demonstrated that intact 
vesicles are not being engulfed by cells in a process resembling phagocytosis. 

Ebert and associates have been engaged in fundamental studies of immune 
mechanisms. The immune response to foreign antigen is an orderly series of 
defensive actions. These include lymphoid cell recognition of the antigen as 
foreign, differentiation of lymphocytes into cytotoxic (killer) cells, proliferation 
of these cells, stimulation of humoral antibody production, and, at the appropriate 
time, suppression of the immune response, which otherwise could get out of hand. 
Thymus-derived lymphoid cells (T cells) play a primary role in this sequence. 

Last year, Ebert and associates described a useful biological system for 
studying the behavior of lymphocytes in the immune response. This consisted 
of a population of T cells exposed to different mitogens, one of which was 
Concanavalin A (ConA). With this tool it was possible to assemble evidence 
supporting an affirmative answer to such questions as: Do lymphocytes differ- 
entiate into killer cells for all antigens by the same mechanism? If so, is the 
sensitization by specific alloantigens simply a triggering of the lymphocyte 
proliferation and concomitant transformation? 

This year Ozato, Ebert, and W. H. Adler, a Visiting Investigator, have con- 
tinued to study T cell recognition. They also have established conditions under 
which suppression of the immune response can be studied. In their experimental 
system, soluble ConA had a suppressor effect. So did spleen cells and lymph node 
cells "coated" with ConA. Contrary to expectation, however, "coated" thymocytes 
were inactive. Modification of the cell surface by ConA was assumed to be the 
immediate cause of the observed differences. This hypothesis was supported by 
the results of studies by Ozato and Ebert using further modifications of lympho- 
cyte surfaces. 

Studies by Ozato, Huang, and Ebert are now going forward on the nature of 
binding of mitogens to lymphocytes in an attempt to delineate the relation 
between cell surface events and subsequent DNA synthesis. 



The activities of this group are largely carried out in Southern California by 
faculty, graduate students, and postdoctoral fellows associated with California 
Institute of Technology. A key individual is Roy J. Britten, who is a staff mem- 
ber of the Carnegie Institution and a Senior Research Associate of Caltech. 

The Developmental Biology Research Group was established as a result of a 
scientific collaboration starting in 1967 between Britten (at the time a member 
of the Biophysics Section of the Department of Terrestrial Magnetism) and Eric 
Davidson (then at Rockefeller University). Each had independently inferred 
that the repetitive DNA sequences, which are ubiquitous in higher organism 
DNA, were involved in the regulation of genetic activity. 

By 1969 Britten and Davidson were convinced that an extensive experimental 
program would be required if their concept of the regulation of genetic activity 
were to be properly tested. They concluded that a large group would be required 
with the potentiality not only to investigate the sequence structure of DNA but 
also to prepare pure and well-characterized RNA from well-defined stages of 
development and to make use of these molecules for subtle studies of their 
sequence relationship and spatial relationship of the sequences in the DNA. 
When Davidson was offered an appointment at Caltech, the decision was made 
to set up laboratories there. 

The work of the new group was initiated late in the spring of 1971 at Pasadena 
and early in the fall at the nearby Kerckhoff Marine Laboratory. Much of the 
effort of the group has been devoted to a thorough investigation of the sequence 
organization of the genome of higher organisms. The rationale offered by Britten 
and Davidson for this choice was : (1) The process which underlies all the other 
levels of control and regulation in the cell must be the control of the expression 
of the genetic information present in the DNA. (2) Thus the control of the 
transcription of the DNA sequences of structural genes into messenger RNA 
is probably fundamental. (3) The control of transcription requires molecular 
recognition of each of the many genes coordinately expressed in a given tissue 
or cell state. (4) It therefore appears likely that each set of genes (batteries) 
for which transcription is coordinately initiated at a given stage of development 
should have a battery-specific similar DNA sequence near the structural gene 
sequence itself. 

This year the DNA of 14 species representing several phyla has been tested 
for interspersion of repeated sequences with single-copy sequences. A very 
similar pattern was observed in 12 of the species. Drosophila and honey bee were 
different. In all the other species the majority of the single copy DNA (65%- 
80%) occurs in lengths of 1000-3000 nucleotide pairs separated by short repeats 
with a typical length of 300 nucleotide pairs. It is now known that most genes 
are single-copy DNA sequences and the observed single-copy lengths are those 
expected for structural genes coding for proteins made up of 300-1000 amino 
acid residues. The measurements made this year also show that many structural 
genes have adjacent repeated sequences. 

The concept that repeated sequences regulate gene expression through trans- 
cription control previously led to the prediction that repeated sequences would 
be found adjacent to structural genes in the DNA. This prediction has been 
borne out by the following measurements. Preparations were made of single- 
copy DNA sequences adjacent to repetitive sequences. This DNA was shown 
to be a specific sequence class containing about one-third of the total single-copy 
DNA. Messenger RNA from developing sea urchin embryos was found to 


hybridize almost as effectively (80% or more) with this DNA preparation as 
with whole single-copy DNA. This observation bears out a prediction by Britten 
and Davidson made from their concepts of genetic regulation and also prepares 
the ground for an examination of the specificity of the adjacent repetitive 
sequences in relation to the state of differentiation of the tissues being studied. 

LOSSES . . . 

I regret that I must report the death of Keith S. McHugh on June 7, 1975. 
Mr. McHugh was a trustee of the Institution from December 1950 to September 
1974, when he resigned for reasons of health. He was named trustee emeritus 
in May 1975 at the annual meeting of the board. Mr. McHugh was an active 
participant in the work of the board of trustees, serving on the auditing com- 
mittee since 1952 and as chairman since 1954; he was chairman of the nominating 
committee for two separate terms. He was also at various times a member of 
the finance and executive committees. 

His business experience and his record of public service made Mr. McHugh a 
great asset to the Institution. He was president of the New York Telephone 
Company from 1949 to 1959. In 1959 he left the telephone company to head 
the New York State Department of Commerce. During 1959-1966 he initiated 
a program for attracting commerce and industry to the state, which succeeded 
in reversing the movement of major companies out of New York. 

Mr. McHugh was appointed to 50 commissions, boards, and special committees 
in New York, including the New York Job Development Authority (chairman), 
the planning coordination board of the Governor's Office for Regional Develop- 
ment (chairman), the Science and Technology Foundation (director), the 
Special Cabinet Committee on Civil Rights, the Atomic Energy Coordinating 
Council, the Air Pollution Control Board, and the Water Resources Commission. 
In 1967 he became chairman of the executive committee of Action for Trans- 
portation, Inc., and in 1968 he headed a post- Vietnam committee to plan social 
and economic transition after the war. 

At the national level, Mr. McHugh was deputy administrator of the National 
Recovery Administration and a member of the National Security Resources 
Board. He was awarded Certificates of Appreciation by the Department of 
Defense for his work on the War Production Board and the Board of War 
Communications during World War II. He was also a member of the Com- 
mission on Revision of Armed Services Pay. 

Mr. McHugh was born at Fort Collins, Colorado, on February 25, 1895. He 
graduated from the University of Wisconsin in 1917 with a B.S. degree in 
chemical engineering. During World War I, he was a captain in the U.S. Army 
in France. 

I must also report, with regret, the death of trustee emeritus Alfred L. Loomis 
on August 11, 1975. Mr. Loomis became a trustee of the Institution in December 
1934 and was a member of the board until his resignation in April 1973. In May 
1973 he was designated trustee emeritus. He was a member of the finance com- 
mittee from 1934 to 1971, except for the war years, and a member of the 
auditing committee from 1956 to 1971. He was a valuable adviser to the Institu- 
tion in scientific as well as fiscal matters. 

Successful in three fields — physics, law, and finance — Mr. Loomis was known 


primarily for his scientific achievements. He graduated from Yale University in 
1909 with a bachelor's degree and a reputation as an inventor with great skill 
in mathematics and science. He then attended Harvard Law School, graduating 
cum laude in 1912. By 1915 he was a member of the law firm of Winthrop & 
Stimson in New York. During World War I Mr. Loomis served as a major in 
the army and helped develop the Aberdeen chronograph, a device that measures 
projectile velocity. After the war he became a highly successful investment 
banker and vice president of Bonbright & Company. 

In 1928 he established the Loomis Laboratories, which developed, among other 
things, the electroencephalograph, a device that records electrical impulses trans- 
mitted by the brain. He also worked on the electronic centrifuge for which he 
and Dr. Edmund Newton Harvey of Princeton received the Wetherill Medal of 
the Franklin Institute in 1935. In 1930 he established the Loomis Institute for 
Scientific Research, and in 1933 retired completely from business to pursue his 
scientific interests. 

During World War II, Mr. Loomis headed the radar research division of the 
Office of Scientific Research and Development under the direction of Carnegie 
president Vannevar Bush. Mr. Loomis was a pioneer in the development of loran, 
the electronic navigational aid. For his war work he received the Medal of Merit 
from the United States government and His Majesty's Medal for Service in the 
Cause of Freedom from the government of Great Britain. 

Mr. Loomis was born in New York, New York, on November 4, 1887. He was 
a founder and trustee of the RAND Corporation, an adviser of the Lawrence 
Radiation Laboratories, and a life member emeritus of the Corporation, Massa- 
chusetts Institute of Technology. 

This year three members of the board of trustees have resigned, and the 
Institution feels their loss. They are Charles P. Taft, Eric Ashby, and Patrick 
E. Haggerty. 

The loss of The Honorable Charles P. Taft is a great one, but it is reassuring 
that he will continue his long association with the Institution as a trustee 
emeritus and that we will still benefit from his guidance and friendship. Mr. 
Taft was a trustee of the Institution from December 1936 to April 1975 — one of 
the longest terms ever served by a trustee. That alone makes his trusteeship 
remarkable, but more than that, we will remember and value his faithful service 
on many committees over the years, giving the Institution the benefit of his 
wisdom and experience. 

The Institution will also miss The Rt. Hon. Lord Ashby of Brandon, a trustee 
since May 1967, who resigned in June 1974. Although distance and his duties as 
Master of Clare College and Vice-Chancellor of The University of Cambridge 
prevented Lord Ashby from taking an active part in the committee work of the 
board of trustees, he has been and will continue to be a good friend of the 
Institution and an important link with the British scientific and educational 

Patrick E. Haggerty, chairman of Texas Instruments Incorporated, elected 
in May 1974, was a valuable addition to our board of trustees. I am sorry that 
he found it necessary to resign from the board in June 1975 because of the many 
contributions he could have made to the work of the board and of the Institution. 


. . . AND GAINS 

The most notable of the gains this year is the election of three new trustees 
at the annual meeting of the board in May 1975. They are John T. Connor, John 
Diebold, and William C. Greenough. 

John T. Conner is chief executive officer and chairman of the board of Allied 
Chemical Corporation. He was Secretary of Commerce under President Lyndon 
Johnson from 1965 to 1967, when he joined Allied Chemical. During World War 
II he was general counsel to the Office of Scientific Research and Development, 
and after the war was counsel to the Office of Naval Research and special 
assistant to the Secretary of the Navy for two years. He was awarded the 
Presidential Certificate of Merit for his service in 1948. He joined Merck and 
Co. in 1947 as general attorney and was president and director from 1955 until 
his appointment as Secretary of Commerce. 

Mr. Connor graduated magna cum laude from Syracuse University in 1936 
and earned a J.D. degree from Harvard Law School in 1939. 

John Diebold is chairman of the board of the Diebold Group, Inc., a manage- 
ment consulting firm, and of John Diebold, Inc., a management and investment 
firm; he founded both companies. He is a member of the Council on Foreign 
Relations, the Hudson Institute, and the National Planning Association. In 1963 
he was appointed to the U.S. delegation to the United Nations science conference 
in Geneva. He is a member of two National Science Foundation groups: the 
advisory committee on ethical and human values of science and technology and 
the industrial panel on science and technology. 

Mr. Diebold received a B.S. degree from the U.S. Merchant Marine Academy 
in 1946; a B.A. in economics, with high honors, from Swarthmore College in 
1949; and an M.B.A., with distinction, from Harvard University in 1959. 

William C. Greenough joined the Teachers Insurance and Annuity Association 
(TIAA) in 1941, became its president in 1957, chairman and president in 1963, 
and chairman and chief executive officer in 1967. He has held the same executive 
positions in the College Retirement Equities Fund (CREF) since it was formed 
as an adjunct to TIAA in 1952. He is a member of The Rockefeller University 
Council and a director of the New York Stock Exchange, the Dry Dock Savings 
Bank, and the Turner Construction Company. 

Dr. Greenough received the A.B. degree from Indiana University in 1935, and 
the University presented him with the Distinguished Service Award in 1960 and 
an LL.D. degree in 1965. He earned the M.A. and Ph.D. degrees from Harvard 
University in 1938 and 1949. In 1951 he received the Elizur Wright Award for 
his study A New Approach to Retirement Income. 

The following honors have been awarded to Staff Members during the past year. 

James D. Ebert., Director of the Department of Embryology, was elected 
to the Institute of Medicine of the National Academy of Sciences. He was 
awarded an honorary D.Sc. degree by Indiana University. Dr. Ebert was also 
Regent's Lecturer at the University of California in the spring of 1975. 

Winslow R. Briggs, Director of the Department of Plant Biology, was elected 
to the American Academy of Arts and Sciences. 

Allan R. Sandage of the Hale Observatories received an honorary D.Sc. degree 


from Miami University, Oxford, Ohio. Dr. Sandage was awarded the Catherine 
Wolfe Bruce Gold Medal by the Astronomical Society of the Pacific. 

Donald D. Brown of the Department of Embryology delivered the Jessup 
Lectures at Columbia University. 

C. Stacy French, Director Emeritus of the Department of Plant Biology, 
received an honorary doctoral degree from the University of Goteborg, Sweden. 

Henrietta H. Swope is a former Staff Member of the Hale Dbservatories who 
retired in 1968 but remains engaged in post-retirement studies. This year she 
was honored by Barnard College, which presented her with the Distinguished 
Alumnae Award, and by the University of Basel (Switzerland), which awarded 
her an honorary doctoral degree. 

J. Eric Thompson,* Staff Member in archaeology with the Division of 
Historical Research from 1935 to 1959, was named Knight of the British Empire 
in the New Year's Honours Lists. Sir Eric also received the Order of the Quetzal 
from the Republic of Guatemala. 

* Died September 9, 1975. 




Baltimore, Maryland 

James D. Ebert 

Staff Members 

Donald D. Brown 
Igor B. Dawid 
Douglas M. Fambrough 
Kenneth J. Muller 
Richard E. Pagano 
Ronald H. Reeder 
Yoshiaki Suzuki 


Ronald D. Brown 
Dana Carroll 
Scott Emmons 
Toru Higashinakagawa 
Leaf Huang 
John F. Morrow 
Seigo Ohi 
Tokindo Okada 
Keiko Ozato 
Aileen K. Ritchie 
Masatoshi Takeichi 
Katherine Tepperman 
William B. Upholt 
Peter K. Wellauer 

Assistant Investigator 
Peter J. Stambrook 


Sandra L. Biroc 
Peter Devreotes 
Marc Friedman 
Carol Kaushagen 
Jose Ramirez 
Changlin A. Wey 

Washington, D.C. 


Hatten S. Yoder, Jr. 

Carnegie Institution 
Distinguished Professor 

Elburt F. Osborn 

Emeritus Research Associate 
Emanuel G. Zies 

Staff Members 

Peter M. Bell 
Francis R. Boyd, Jr. 
Felix Chayes 
John R. Cronin 
Gordon L. Davis 
David H. Eggler 
Larry W. Finger 
John D. Frantz 
P. Edgar Hare 
Thomas C. Hoering 
T. Neil Irvine 
Thomas E. Krogh 
Ikuo Kushiro 
Ho-Kwang Mao 
Douglas Rumble III 
David Virgo 


Jagannadham Akella 
Richard J. Arculus 
John I. Hedges 
Floyd N. Hodges 
Frank E. Huggins 
Bruce R. Lipin 
Robert H. McCallister 
Gilford H. Miller 
Karlis Muehlenbachs 
Bj0rn 0. My sen 
Yasuo Nakamura 
Howard R. Naslund 
Yoshikazu Ohashi 
Martin G. Seitz 
Rosemary J. Vidale 
E. Bruce Watson 


Dora Lee 
Noreen Tuross 



Pasadena, California 


Horace W. Babcock 

Associate Director 
J. Beverley Oke 

Staff Members 

Halton C. Arp 
Edwin W. Dennison 
Jesse L. Greenstein 
James E. Gunn 
Robert F. Howard 
Jerome Kristian 
Robert B. Leighton 
Guido Munch 
Gerry Neugebauer 
George W. Preston 
Bruce H. Rule 
Allan R. Sandage 
Wallace L. W. Sargent 
Maarten Schmidt 
Leonard T. Searle 
Stephen A. Shectman 
Arthur H. Vaughan, Jr. 
James A. Westphal 
Harold Zirin 

Staff Associates 

Eric E. Becklin 
Robert J. Brucato 
Michael W. Werner 


Ermanno F. Borra 
Michael H. Hart 
Francois Schweizer 
Pieter van der Kruit 
Christopher P. Wilson 
Hirokazu Yoshimura 
Robert J. Zinn 

Carnegie-Chilean Fellows 

Eduardo Hardy 
Maria Teresa Ruiz 

Frances Cordova 
Jonathan H. Elias 
Richard F. Green 
Paul Hickson 
John G. Hoessel 
John P. Huchra 
John Kormendy 
Philip Massey 
Jorge Melnick 
William C. Priedhorsky 
Douglas M. Rabin 
Anneila I. Sargent 
William L. Sebok 
Edwin L. Turner 
Barry E. Turnrose 
William E. Westbrook 
Theodore B. Williams 
Steven P. Willner 


Stanford, California 


Winslow R. Briggs 

Staff Members 

Joseph A. Berry 
Olle Bjorkman 
Jeanette S. Brown 
David C. Fork 
Malcolm A. Nobs 
William F. Thompson 


C. Stacy French, Director 
William M. Hiesey, Staff Member 


Steven J. Britz 
William G. Hagar 
Tetsuo Hiyama 
John M. Mackenzie 
Bruce E. Mahall 
Norio Murata 
Charles Rogler 
Heather G. Strong 
John H. Troughton 

Student Observers 

Jill Bechtold 

Steven V. W. Beckwith 


Daniel J. Cosgrove 
James Ehleringer 



Washington, D.C. 

George W. Wetherill 

Associate Director 
L. Thomas Aldrich 

Albrecht W. Hofmann 
Bill H. Hoyer 
David E. James 
Alan T. Linde 
Nancy R. Rice 
Richard B. Roberts 
Vera C. Rubin 
I. Selwyn Sacks 
Norbert Thonnard 
Kenneth C. Turner 

Distinguished Service Member 
Merle A. Tuve 


Scott E. Forbush 

Staff Members 

George E. Assousa 
Manuel N. Bass 
Ellis T. Bolton 
Louis Brown 
Dean B. Cowie 
W. Kent Ford, Jr. 
Stanley R. Hart 


John Bannister 
Charles L. Bennett 
Arturo Cuyubamba 
Michael B. Davis 
Charles Doering 
Antonio Flores 
George H. Pepper 
Charles J. Peterson 
J. Arthur Snoke 
John W. Warner III 
William M. White 


David Westpfahl, Jr. 

Reports of Departments 
and Special Studies 

Department of Embryology 

Department of Terrestrial Magnetism 

Hale Observatories 

Geophysical Laboratory 

Developmental Biology Research Group 

Department of Plant Biology 

Department of Embryology 

Baltimore, Maryland 

James D. Ebert 

Carnegie Institution of Washington Year Book 74, 1974-1975 


Introduction 7 

Developmental Genetics by Gene Isolation: The Dual 5S DNA System 12 

Purification and characterization of new 5S DNAs from 

Xenopus laevis and Xenopus mulleri 13 

Nucleotide sequences adjacent to the 5S RNA genes 13 

The arrangement of length heterogeneity in 5S DNA of Xenopus laevis 14 

Transcription of 5S DNA 19 

Xenopus laevis-Xenopus mulleri hybrid animals 19 

The Molecular Basis, Arrangement, and Inheritance of Length 

Heterogeneity in Xenopus laevis Ribosomal DNA 19 

The organization of sequences within the spacer of Xenopus laevis ribosomal DNA 20 

The arrangement of length heterogeneity in amplified and chromosomal rDNA . 22 

The inheritance of rDNA spacer length heterogeneity 25 

Conclusion - 26 

Chromosomal Proteins and the Control of Gene Function 26 

The association of ribosomal genes with histones in cultured cells 27 

Analysis of the proteins of isolated ribosomal gene chromatin 28 

Iodination of Xenopus histone F2ai in chromatin 33 

The Regulation of Fibroin Genes 36 

Morphological observations of silk gland nuclei during development 36 

Preparation of assay probes for detecting fibroin mRNA synthesized in vitro 38 

Fibroin mRNA synthesis in a pulse-labeled silk gland 39 

Fibroin mRNA synthesis in isolated silk gland nuclei 41 


Localization of Fibroin Genes in Silk Gland Cells of Bombyx mori 42 

Attempts to Isolate the Fibroin Gene of Bombyx mori 45 

Biogenesis of Mitochondria 46 

Mapping studies on mitochondrial DNA of Xenopus laevis and Xenopus m,ulleri . . 46 

Physical mapping of 4S RNA genes on mitochondrial DNA 47 

A comparative study of the mitochondrial genomes of the sheep and the goat .... 48 

Mitochondrial DNA and RNA in Drosophila 49 

Phospholipid-Cholesterol Dynamics in Mammalian Cells 53 

Interactions of phospholipid vesicles with mammalian cells 54 

Studies of mechanism 54 

A proposed mechanism for vesicle-cell interactions 59 

Interaction of phospholipid vesicles with mouse lymphocytes in vitro 60 

Spatial distribution of cell surface lipids 64 

Intercellular exchange of lipids 65 

Lipid perturbations and lymphocyte activation 66 

Studies on Skeletal Muscle Plasma Membranes 67 

Characterization of receptor pools 68 

Turnover of receptors in denervated adult muscle 74 

Iodination of membrane proteins of muscle cells 78 

An investigation of the involvement of denervation in the human skeletal muscle 

diseases amyotrophic lateral sclerosis and congenital myotonic dystrophy .... 81 

Signaling in the Nervous System : Synaptic Morphology of 

Specific Neurons in the Leech 85 

Mouse Lymphocytes as Tools for the Study of Mitogenic Events and 

Regulatory Processes Associated with the Cell Membrane 88 

Regulation of transplantation immunity by T lymphocytes 88 

Modification of the lymphocyte surface by Con A 90 

The study of ConA binding sites on murine thymic lymphocytes 93 

Enhancement of surface permeability of thymocytes by ConA 94 

The Collection of Human Embryos 97 

Developmental stages in human embryos 97 

Development of the nervous system 98 

Staff Activities 98 

Bibliography 101 

Personnel 103 


On an earlier occasion I wrote, in part, places the entire scientific enterprise in 

"In recent years it has become increas- jeopardy. As Hershey put it, "duplica- 

ingly fashionable for directors of labora- tion of effort — is a virtually nonexistent 

tories to stress that their organizations bugbear." 

are 'where the action is,' with increasing This lesson, which must be restated 
emphasis on the interactions of society repeatedly, is brought home by a look 
and technology underlying our social at Donald Brown's present interests. For 
concerns." This posture is fully conso- several years they were generated almost 
nant not only with currently increasing exclusively by his technical ingenuity, 
emphasis on contract-supported, directed single-mindedness, and tenacity, and by 
research and the decline in grants for his own belief that it should be possible 
unfettered basic investigations, but with to isolate and characterize the genes of 
today's cry for the dissolution of peer higher organisms, and attain an even- 
review mechanisms. According to current tual understanding of the mechanisms 
dogma, self-serving scientists are in- regulating their action. Still further ad- 
capable of assessing the long-term soci- vances are recounted in the pages to 
etal consequences of their research. The follow, but it is now possible to see more 
Department of Embryology has been clearly how crucial to these new direc- 
privileged to play a major role in re- tions have been major findings in other 
search that has provided the under- laboratories on two fronts: the use of 
pinning for advances in biology and restriction enzymes for elucidating the 
medicine which have had a significant molecular fine structure of genes; and of 
impact in promoting human welfare. I the newer methodology for covalently 
have addressed this question on other linking DNAs of different organisms; for 
occasions. It will suffice to say here that example, fragments of DNA from the 
we have not neglected our societal re- frog, Xenopus laevis, or the silkworm, 
sponsibilities. Yet when I hear the ex- Bombyx ?nori, may be linked to a bac- 
pression, "where the action is," I still terial plasmid, enabling the replication 
think not of technology but of basic sci- of animal DNA sequences in bacteria, 
ence, which we need today more than I would be less than honest if I did 
ever before. not say that almost without exception 

At a time when basic science is ques- the staff, fellows, and students in the 

tioned, even ridiculed, how can we our- Department have enjoyed a productive 

selves best understand, and best explain year. This "admission," however, makes 

to others, the guiding spirit of our work, my next task well-nigh impossible. How 

which as Alfred Hershey once wrote, does one extract from such an impressive 

". . .is scarcely evident in a day to day record of accomplishment and happy 

record of it." I believe, with Hershey, associations the year's "highlights"? One 

that we must repeatedly stress the com- has to admit further that he is guided 

munal nature of scientific effort. Im- in part by his personal taste ; but beyond 

portant gains made in our Department that I have chosen to point up two 

have contributed to advances made by classes of findings — those in which an- 

others; all too often, however, we fail swers have been provided to specific 

to stress the obverse truism, that im- questions, and those which have in- 

portant conceptual and technical gains creased the number of valid questions, 

made in laboratories other than our own New insights into gene structure have 

are vital to our own progress. It is in this emerged through the use of restriction 

light that we may best give the lie to enzymes. Unexpected features of the 

the argument that duplication of research structure of the ribosomal genes have 



been discovered recently by a group in- 
cluding postdoctoral Fellow Peter Wel- 
lauer and Staff Members Ronald Reeder, 
Igor Dawid, and Donald Brown, who 
have been studying the "anatomy" of 
the DNA which codes for 28S and 18S 
ribosomal RNA (termed ribosomal DNA 
or rDNA). In Xenopus laevis, about 450 
copies of these genes are present on each 
of two chromosomes in every cell. Earlier 
studies revealed that these genes can be 
isolated completely free of the rest of 
the cell's DNA. Work in the Department 
and elsewhere has shown that rDNA is 
composed of multiple repeating units 
tandemly linked end to end. About two- 
thirds of each repeating unit is a gene 
region which serves as the template for 
the transcription of 28S and 18S ribo- 
somal RNA. The remaining one-third 
of each repeating unit is not transcribed 
into RNA, has no known function, and 
is called the spacer region. 

Within a cell, all 900 copies of the 
transcribed gene region appear to be 
identical both in length and in nucleotide 
sequence, suggesting that cells have some 
mechanism for maintaining these hun- 
dreds of copies identical within close 
tolerances. The recent, unexpected find- 
ing is that the spacer regions between 
the genes come in different lengths. As 
many as seven different spacer sizes have 
been found on a single chromosome alter- 
nating with gene regions of constant 

The existence of heterogeneous spacer 
lengths in rDNA was discovered by the 
use of a site-specific restriction endo- 
nuclease called EcoRI. EcoRI is one of 
a newly described class of enzymes 
which recognize specific short nucleotide 
sequences in the DNA and will cut the 
DNA only at the site of those sequences. 
In rDNA the EcoRI recognition site 
occurs just twice in each repeating unit. 
Cutting the DNA at these two sites pro- 
duces two fragments per repeating unit 
— one fragment containing only gene se- 
quences and the other containing all of 
the spacer plus some gene region on each 

end. Electrophoresis of these fragments 
on agarose gels separates them according 
to size and reveals that the spacer- 
containing fragments come in different 
size classes while the gene fragment is 
homogeneous in size. 

When EcoRI is used to cut rDNA 
isolated from individual frogs, it is found 
that the pattern of spacer length can 
vary from one individual to the next. 
This has made it possible to use the 
rDNA spacer pattern as a genetic 
marker, that is, to select two frogs which 
have different spacer patterns, mate 
them, and observe how the spacer pat- 
terns are transmitted to the progeny. 

Dana Carroll has used a similar ap- 
proach in further investigating the na- 
ture and organization of heterogeneity 
in the oocyte-type 5S DNA of Xenopus 
laevis. Again, not only is the existence 
of heterogeneity confirmed, but it is 
shown to be located in the AT-rich por- 
tion of that spacer. 

In these examples, at least, length 
heterogeneity results from the presence 
of variable amounts of reiterated simple 
sequences in nontranscribed spacer, with 
the region of variable length being close, 
if not adjacent, to the 3' end of the 
transcription unit. 

It is hoped that experiments of this 
sort will eventually reveal the mecha- 
nism whereby a cell can conserve 
hundreds of apparently identical gene 
regions while allowing them to be inter- 
spersed with heterogeneous, less con- 
served spacer regions. 

Carroll's findings are a part, albeit a 
key part, of a larger undertaking by 
Brown and his colleagues, namely the 
exploration of the dual 5S DNA system 
of Xenopus. The Xenopus genome con- 
tains at least two kinds of 5S DNA. 
"Somatic" 5S DNA appears to be made 
in all cells. "Oocyte" 5S DNA synthesis 
is "turned off" in somatic cells. Key 
questions concerning gene regulation are 
then posed. Donald Brown and E. Jordan 
describe further progress in isolating 5S 
DNAs; and in a study related closely to 


Carroll's, Ronald Brown has analyzed in number and distribution of acetylcholine 

detail the spacer sequence adjacent to receptors in the plasma membrane in 

the 3' end of the 5S DNA transcription various physiological states, their rate of 

unit. It is thought that the end of the synthesis and destruction, and the mecha- 

gene itself and the first two spacer nism by which they are incorporated into 

nucleotides may be at least part of the the plasma membrane of muscle fibers 

termination signal for RNA polymerase, and later removed and degraded by the 

Significant progress is again reported fibers, 

by Douglas Fambrough and his col- Fambrough and his co-workers have 

leagues, especially graduate student continued to pursue their studies on the 

Peter Devreotes, postdoctoral Fellow mechanisms of synthesis, incorporation 

Katherine Tepperman, and Fambrough's into plasma membrane, and degradation 

collaborator at the Johns Hopkins Uni- of acetylcholine receptors. Their working 

versity School of Medicine, Daniel hypothesis includes the following ele- 

Drachman. ments. Acetylcholine receptors are syn- 

For several years the group has been thesized by the normal protein synthesiz- 
focusing on the differentiation of the ing machinery, this synthesis being sensi- 
skeletal muscle plasma membrane and tive to inhibition by puromycin and by 
the organization of skeletal muscle fibers, cycloheximide. Newly synthesized re- 
in previous Year Books they have re- ceptor components occur as organized 
ported not only on the appearance and macromolecular structures of molecular 
maturation of several plasma membrane weight about 250,000 daltons (when 
properties related to the specialized func- solubilized in detergent solution) in- 
tions of skeletal muscle during develop- corporated into internal cell membranes 
ment but also on some of the changes in (of unknown type and location). These 
plasma membrane properties which char- precursors are then inserted into the 
acterize denervated adult skeletal muscle plasma membrane, the population of pre- 
and are reminiscent of the properties of cursors being enough to support two 
embryonic muscle. They have also re- hours of new receptor incorporation into 
ported on some characteristics of abnor- plasma membrane without any new pro- 
mal muscle, namely in the muscular tein synthesis. Once in the plasma mem- 
dysgenic mouse and in humans with the brane the receptors are functional mole- 
disease myasthenia gravis. Many of their cules and are free to interact with the 
studies have focused upon the acetylcho- radio-labeled probe, iodinated a-bungaro- 
line receptors of skeletal muscle, since toxin. Plasma membrane receptors are 
the receptors appear in the plasma mem- degraded by a random-hit process which 
brane as readily detected functional is energy dependent and proteolytic, 
molecules early in muscle development, probably involving internalization of 
they are later confined to a small area membrane containing receptors and then 
of cell surface — the postsynaptic surface degradation by lysosomal enzymes. Dur- 
of the neuromuscular junction — and in ing the past year the group has developed 
denervated skeletal muscle they appear techniques which will allow direct testing 
all over the surface of each muscle fiber, of several aspects of this model of the 
Because acetylcholine receptors can be life history of receptor molecules. They 
readily identified not only by their func- have also devised a method for contin- 
tion but also through the use of a radio- uous monitoring of receptor degradation 
active probe which binds to them with in organ-cultured skeletal muscle and 
extreme tenacity (iodinated a-bungaro- have begun extending their findings to 
toxin), they can be studied in ways that adult mouse muscle, 
are unique. Thus the group continues to Drachman and Fambrough have also 
understand more each year about the investigated another human disease, con- 


genital myotonic dystrophy, in which the and the generation of cytotoxic lympho- 
muscle pathology suggested the possible cytes was studied in congenic-resistant 
participation of denervation. Since the mice using cortisone-resistant thymo- 
appearance of extrajunctional acetyl- cytes (CRT) as responding cells'. LPS 
choline receptors is a sensitive marker enhanced the generation of cytotoxic 
for denervated skeletal muscle surface, lymphocytes selectively when subopti- 
they have used the radio-labeled probe mal concentrations of alloantigens were 
for acetylcholine receptors on biopsy given followed by a selective augmenta- 
specimens from patients with congenital tion in the MLR. Mitogenic concentra- 
myotonic dystrophy. They have also tions of ConA, on the other hand, sup- 
examined biopsy material from two pa- pressed the generation of cytotoxicity 
tients with the denervating disease amyo- regardless of antigen dose. The mecha- 
trophic lateral sclerosis. nism of suppression could not be ascribed 

During 1974-1975 Richard Pagano was to suppressor T cells, since the addition 
joined by two postdoctoral Fellows, Leaf to the culture of syngeneic CRT acti- 
Huang and Masatoshi Takeichi. To- vated by ConA did not change the im- 
gether with Ebert's colleague, Keiko mune response. Prospective suppressor 
Ozato, they have vigorously extended cells that can be activated by ConA 
their explorations of phospholipid-cho- were found to be located in lymphoid 
lesterol dynamics in mammalian cells, organs other than thymus, because sup- 
Their objective is to understand the dy- pression occurred in CRT mixed lympho- 
namic movements which cellular phos- cyte cultures by ConA-activated spleen 
pholipids and cholesterol can undergo in cells or ConA-activated lymph node cells, 
cell membranes, and the effects which These cells also suppressed MLR. Sup- 
perturbations in these motions can have pressor activity was resistant to mito- 
on cellular activity. Among their studies mycin C treatment and sensitive to 
described in the body of the Report, two anti-0 antibody plus complement. ConA- 
are especially noteworthy. They have activated thymocytes, on the contrary, 
combined their earlier knowledge of in- amplified the proliferation of the re- 
teractions of cell membranes and lipid sponding cells. 

vesicles with the technique of high reso- In an attempt to modulate the recog- 

lution electron microscope autoradiog- nition processes that occur on lympho- 

raphy to determine the surface distribu- cyte membranes in mixed lymphocyte 

tion of one lipid component, phosphatidyl cultures, responding cortisone-resistant 

choline, on Chinese hamster fibroblasts, thymocytes or stimulating spleen cells 

Second, in cooperation with Ozato and (treated with mitomycin C) were pre- 

Ebert, they have begun to examine the treated with native ConA (N-ConA) or 

effects of cell surface lipid perturbations succinyl ConA (S-ConA). Highly sig- 

on lymphocyte activation. nificant cell proliferation was observed 

Ozato and Ebert, with the collabora- in syngeneic combinations when either 

tion of William Adler of the Gerontology the responding cells or the stimulating 

Research Center, National Institutes of cells were treated with ConA, although 

Health, have continued to focus on ConA alone was never mitogenic. In 

mouse lymphocytes in their studies of allogeneic combinations the proliferative 

mitogenic events regulated by processes response on day 3 was five to seven times 

associated with the cell membrane. higher than in the normal MLR by the 

Two significant discoveries have ConA pretreatment of either partner, 

emerged. First, the regulation by two The triggering of proliferation was de- 

mitogens, Concanavalin A (ConA) and pendent on two factors : (1) The presence 

bacterial lipopolysaccharide (LPS), of of spleen cells as the stimulating cells 

the mixed lymphocyte reaction (MLR) (thymocytes were much less effective). 



(2) The presence of ConA molecules (the 
effect was blocked by the specific in- 
hibitor of ConA, aMannopyranoside) on 
either one of the partners. The optimal 
concentration of S-ConA was about twice 
that of N-ConA. Even more striking was 
the observation that cultures in which 
either one of the partners was pretreated 
with ConA in allogeneic combinations 
showed a strong suppression (60% to 
80% inhibition) in the subsequent gen- 
eration of the cytotoxic lymphocytes 
(CL). The ConA concentration required 
to trigger proliferation corresponded to 
that for suppressing the generation of 
CL. ConA pretreatment did not exhibit 
cytotoxic activity toward syngeneic 
tumor cells. 

It is a special pleasure to report the 
arrival, as this report is being written, 
of Staff Member Kenneth J. Muller. 
Muller, trained in neurobiology at Mas- 
sachusetts Institute of Technology and 
Harvard Medical School, comes prepared 
to devote himself, for the foreseeable 
future, to the study of synaptic trans- 
mission and integration in the central 
nervous system, using, at least at the 
outset, the leech Hirudo medicinalis as 
his experimental object. His brief state- 
ment, "Signaling in the nervous system: 
Synaptic morphology of specific neurons 
in the leech," based on work performed 
jointly with U. J. McMahan of Harvard 
Medical School, will serve to introduce 
him and his subject. 

Mention has already been made of 
the contributions of several Fellows: 
Peter Wellauer, having completed his 
third year in the Department (in 1974- 
1975 under the auspices of the National 
Cystic Fibrosis Research Foundation), 
will spend most of 1975-1976 at Cali- 
fornia Institute of Technology before 
returning to his native Switzerland; late 
in the spring of 1975, Dana Carroll 
moved to the Department of Microbiol- 
ogy at the University of Utah School of 
Medicine; Ronald Brown will join the 
University of Cincinnati. Fellows Kath- 
erine Tepperman, Leaf Huang, Masa- 

toshi Takeichi, and Keiko Ozato will 
continue their associations with the De- 
partment in 1975-1976, the first two as 
Fellows of the U.S. Public Health Serv- 
ice; Takeichi, as a Carnegie Fellow; and 
Ozato, supported jointly by Carnegie 
and the Marine Biological Laboratory. 

In addition to Ronald Brown and 
Dana Carroll, two other postdoctoral 
Fellows have been associated with Don- 
ald Brown: John Morrow and Scott 
Emmons, both Fellows of the National 
Cystic Fibrosis Research Foundation. 
Morrow, who has already taken up new 
responsibilities at Harvard Medical 
School, reports some progress in his at- 
tempts to isolate the fibroin gene of 
Bombyx mori. Emmons is trying to de- 
velop an assay system which can be used 
to examine the transcription products 
synthesized using pure 5S DNA as a 
template in vitro. 

In addition to Wellauer, postdoctoral 
Fellows in Dawid's program include 
Seigo Ohi and William Upholt. Upholt 
has centered his attention on comparing 
the mitochondrial genome of the sheep 
and goat. He is leaving (August 31, 
1975) to join the research staff of the 
University of Chicago. Ohi, whose ap- 
pointment is also drawing to a close, has 
continued to pursue an important tech- 
nical objective, the covalent attachment 
of ferritin to 4S RNA for the visualiza- 
tion of 4S RNA/DNA complexes in the 
electron microscope. 

Also completing a two-year stay is 
Toru Higashinakagawa, a joint Carnegie- 
Mitsubishi Fellow. He will return to his 
post at the Mitsubishi-Kasei Institute of 
Life Sciences in Tokyo. 

Aileen Ritchie has continued her stud- 
ies with Douglas Fambrough as a Fellow 
of the Muscular Dystrophy Associations 
of America, Inc. 

In the summer and fall of 1974 and 
again in the spring of 1975 the Depart- 
ment was privileged to have the company 
(all too briefly) as a Senior Carnegie 
Fellow of Professor Tokindo Okada of 


Kyoto University, with whose laboratory Kaushagen, J. L. Ramirez (both with 

the Department has long enjoyed a Dawid), Sandra Biroc (with Reeder), 

highly productive association. and, as cited above, Peter Devreotes. 

Extramural collaborators continue to Ms. Biroc has accepted a postdoctoral 

play an important role in the research of appointment at the University of Cali- 

the Department. This year the group fornia, Davis. 

included W. H. Adler, Daniel Drachman, Mentioning Davis, California, takes 

and U. J. McMahan, already mentioned; us naturally to the Carnegie Embryo- 

A. Forsheit and N. Davidson of the logical Collection, which was moved 

California Institute of Technology ; there in July 1973. The Collection is now 

George Brownlee of the MRC Unit of housed, as Ronan O'Rahilly and Ernest 

Molecular Biology, Cambridge; and Gardner put it, in "Carnegie Village," 

Hayden Coon of the National Cancer in a delightful setting halfway between 

Institute. the School of Medicine and the Cali- 

Graduate students continued to be an fornia Primate Research Center. Work 

integral part of the Department's life, is already well under way, but a "for- 

Those whose contributions are presented mal" opening is being planned for No- 

specifically in the Report are Carol vember 1975. 



D. D. Brown, R. D. Brown, D. Carroll, S. Emmons, and E. Jordan 

Much of our attention is now focused part. The X. laevis 5S DNA contains 

on the dual 5S DNA system of Xenopus. only oocyte-type genes ; the purified X. 

Our long-term goals are to purify the mulleri 5S DNA has been identified as 

two kinds of 5S DNA from the genomes oocyte-type 5S DNA as well, 

of X. laevis and X. mulleri (and perhaps We describe progress toward our goal 

other species of Xenopus as well), to of isolating other 5S DNA components 

sequence regions in their spacer DNAs from the Xenopus genome. Dana Carroll 

which may have regulatory functions, has demonstrated that 5S DNA "repeats" 

and to reconstruct their transcriptional of different lengths can be adjacent in 

controls in vitro. the genome, a finding similar to that 

Xenopus synthesizes at least two kinds described for rDNA in another section 
of 5S ribosomal RNA. Somatic cells pro- of this Report (P. K. Wellauer et al., this 
duce one type, termed somatic 5S RNA. Report). Ronald Brown has analyzed 
The predominant 5S RNA synthesized the spacer sequence adjacent to the 3' 
by oocytes differs from the somatic 5S end of the 5S DNA transcription unit. 
RNA by six nucleotides. However, some The AT-rich spacer begins immediately 
somatic 5S RNA as well as small amounts at the 3' end of the gene. The end of the 
of other 5S RNAs are synthesized in gene itself and the first two spacer nu- 
oocytes. Thus a control mechanism exists cleotides are proposed to be at least part 
which appears to shut off oocyte 5S RNA of the termination signal for RNA poly- 
genes in somatic cells. In previous Re- merase. Scott Emmons has begun to de- 
ports we have described the isolation of velop an assay system which can be used 
5S DNAs (the DNAs coding for the 5S to examine the transcription products 
RNAs) from X. laevis and X. mulleri. synthesized using pure 5S DNA. as a 
These DNAs have been sequenced in template. 



Purification and Characterization of 

New 5S DNA from Xenopus laevis and 

Xenopus mulleri 

D.D. Brown, E. Jordan, R.D. Brown, 
and D. Carroll 

The purified 5S DNAs we have been 
studying for several years code for 
oocyte-type 5S RNA. We have predicted 
that somatic 5S RNA will be transcribed 
from a completely different DNA com- 
ponent. These genes should be present 
in multiple copies and band as satellite 
DNA. We have detected and purified a 
new DNA component containing 5S 
RNA genes from the genome of each of 
the two Xenopus species we have been 
using. The new 5S DNAs contain about 
10% as many 5S RNA genes per cell as 
do the oocyte 5S DNAs. The new X. 
laevis 5S DNA has been purified and 
characterized in some detail; the new 
X. mulleri 5S DNA is still only about 
60% pure and has not been characterized. 

The new X. laevis 5S DNA is purified 
by four density gradient steps. It has 
been characterized by thermal denatura- 
tion, analytical ultracentrifugation, dena- 
turation mapping, and restriction enzyme 
digestion. Table 1 compares character- 
istics of the new 5S DNA component 
with those of oocyte-type 5S DNA from 
X. laevis. 

The spacer regions of the two 5S 
DNAs have little if any homology by 
hybridization criteria. Moreover, studies 

of the new 5S DNA demonstrate that 
the gene sequence is not oocyte-type. 
Although it contains some sequences 
predicted for the somatic-type 5S RNA, 
the new 5S DNA has some different 
sequences as well. Our tentative con- 
clusion is that we have isolated a 5S 
DNA component containing a gene which 
codes for a 5S RNA different from both 
the oocyte-type and the somatic-type 
5S RNAs. 

A new 5S DNA from X. mulleri has 
been partly purified. It is a trace com- 
ponent of the genome, containing only 
a small fraction of the number of 5S 
RNA genes compared to the oocyte-type 
5S DNA. Sequence studies have not yet 
been carried out. 

Nucleotide Sequences Adjacent to 
the 5S RNA Genes 

R.D. Brown 

The sequences adjoining the 3' end of 
the genes coding for oocyte-type 5S RNA 
from X. laevis and X. mulleri have been 
compared. 5S RNA was hybridized to the 
coding strand of the 5S DNA. This 
hybrid was used as a template for E. 
coli DNA polymerase I, utilizing two 
important properties of this enzyme. 
First, the enzyme requires a primer for 
nucleic acid synthesis. It will initiate 
synthesis on the hybrid only by adding 
nucleotides to the 3' terminus of the 5S 
RNA. Second, in the presence of Mn+ + 

TABLE 1. Characteristics of the Newly Isolated 5S DNA from X. laevis 
Compared to Oocyte 5S DNA 

Oocyte-5S DNA 

New 5S DNA 

Buoyant density in CsCl 

Strand separation in alkaline CsCl 

Repeat length (base pairs) 

Approximate number of repeats/haploid DNA 

% 5S DNA of total DNA 

Hind III 
Hae III 

Type of gene 











r ity (cleavages/repeat) 







oocyte-5S RNA 

unknown 5S RNA 


individual ribonucleotides may be poly- the 3' proximal sequence, rCrUrUTTCA 

merized into the growing DNA chain AArG(T), plus the oligonucleotides 

along with the other three deoxyribonu- TTTTCAAArG(T) and T M G(A). These 

cleotides. This "ribosubstitution" intro- sequences are essentially identical to por- 

duces sites into the DNA which may be tions of those determined by Brownlee 

cleaved specifically with alkali or RNase. and Brown {Year Book 73, p. 12) to be 

The first few nucleotides adjacent to the AT-rich portion of the 5S DNA 

each gene were sequenced by the analysis spacer region. In Fig. 1, they are aligned 

of limited synthesis with only one or two with one of the most common sequences 

deoxynucleotides present in the reaction that derive from this portion of the 

mixture. The terminal sequence of X. spacer. We conclude that the AT-rich 

laevis and X. mulleri 5S RNA is region of the X. laevis spacer is immedi- 

GCUUoh- Limited synthesis led to the ately adjacent to the 3' end of the gene 

conclusion that the sequences added to (see also Fig. 3). 
this terminus were TTCA on the X. 
laevis 5S DNA and TT(A or C)G on the 

X. mulleri template. At the third nucleo- The Arrangement of Length 
tide after the gene, the two spacer se- Heterogeneity in 5S DNA of 
quences diverge and the X. mulleri Xenopus laevis 
sequence is heterogeneous. We assume D Q ano r, 
that sequences important for the termi- 
nation of RNA synthesis are likely to be We have been investigating the nature 
conserved in the two Xenopus species, and organization of heterogeneity in the 
If this is true then the termination signal oocyte-type 5S DNA of X. laevis. Last 
must end two nucleotides after the gene year, we reported the existence of repeat 
and is likely to contain at least the two length heterogeneity, which was discov- 
terminal nucleotides of the gene sequence ered by analysis of Hind III restriction 
itself: (gene) 5'-TTTT-3' enzyme digests of 5S DNA. We have 
3'-AAAA-5' (spacer). confirmed the existence of this hetero- 

The heterogeneity of the X. mulleri geneity and its location in the AT-rich 

sequence adjacent to the gene has pre- portion of the spacer by examining diges- 

vented further sequencing of this region, tion products of the 5S DNA produced 

It has been possible to show, however, by another enzyme, Hae III. This enzyme 

that the nucleotide following the first cleaves DNA at the sequence 5'-GGCC- 

G residue is heterogeneous and may be 3'. There are two of these sequences 

T, A, or C and possibly also G. The X. within the 5S RNA gene, between posi- 

laevis sequence in this region has been tions 8 and 9 and positions 66 and 67. 

more readily analyzed. Short extension When the actual digest is analyzed by 

products, ribosubstituted with rGTP and gel electrophoresis (Fig. 2) , four frag- 

digested with T x RNase, contain at least ments are observed, so the enzyme also 

3 ' END OF 

" C-A-A-A-G-T-T-T-G-A-G-T-T-T-T 

5' C-U-U-T-T-C-A-A-A-G(T) T n G (A) T-T-T-T-C-A-A-A-G (T) 3' 

2 16 3 

Fig. 1. Spacer sequences adjacent to the 3' end of the 5S rRNA gene in X. laevis. Three 
oligonucleotides adjacent to the gene are aligned with one of the most abundant spacer 
sequences found previously within the A-T rich spacer region (see Year Book 73). The 
previous sequencing was done on the strand opposite to the one drawn here. 



Fig. 2. Electrophoresis in a 6% polyacryla- 
mide gel of X. laevis 5S DNA digested with 
Hae III (left) and Hae III + Hind III (cen- 
ter). A sample of bacteriophage 0X174 RF 
DNA digested with Hae III is included (right) 
for molecular weight standards. Electrophoresis 
is from top to bottom. 

Hind HI 


A + T 
1 l I » I 1 1 I 


5' 120 3' 

makes two cuts in the spacer. The small- 
est fragment is 58 base pairs long, the 
presumed intragenic piece. A homogene- 
ous fragment of 180 base pairs and a 
heterogeneous collection of DNA mole- 
cules averaging approximately 450 base 
pairs are generated. These heterogeneous 
fragments separate on gels into distinct 
sub-bands with the same mass distribu- 
tion and average sub-band separation 
(14 base pairs) observed for the repeat- 
length fragments generated by Hind III. 

Co-digestion of 5S DNA with Hae III 
and Hind III (Fig. 2) yields the 58 and 
180 base-pair fragments unchanged. The 
~450 base-pair collection of hetero- 
geneous length molecules are shortened 
to ~400 base pairs, and a new homo- 
geneous fragment of 45 base pairs ap- 

Our current model of X. laevis 5S 
DNA (Fig. 3) comes from the restriction 
enzyme data and nucleotide sequencing 
(see the preceding report by R. D. 
Brown). The length heterogeneity resides 
entirely within the AT-rich spacer; re- 
peat units appear to differ from each 
other by the number of internal (15- 
mer) repeats they contain. The arrange- 
ment of this heterogeneity will help to 
distinguish between sudden and gradual 
mechanisms of tandem gene evolution 
(P. K. Wellauer et al., this Report). 

We have investigated the possibility 
that some of the heterogeneity might be 
due to differences between animals, as 
observed in the case of the ribosomal 
DNAs. 5S DNA was isolated from four 
frogs, and Hind III digests of each were 
analyzed by gel electrophoresis. The 5S 


l I 1 1 i i 1 1 1 1 1 1 i 1 1 i 1 1 1 1 1 1 1 i i l 1 1 i 

Mae IK 


1 1 

G + C 


Fig. 3. A model of the repeating unit of X. laevis 5S DNA, indicating the location of the 
cutting sites for Hind III and Hae III. 



DNA purified from individual animals 
was as heterogeneous in repeat lengths 
as the pooled 5S DNA. 

To analyze the heterogeneity of 5S 
DNA repeat lengths along single mole- 
cules, we joined fragments of 5S DNA 
with a bacterial plasmid DNA. We dis- 
covered first that the tetracycline-re- 
sistant plasmid, PSC101, has a single 
Hind III site, very close to its EcoRI 
site, which had been used for the inser- 
tion of foreign DNA. We then demon- 
strated that stable insertions could be 
made at the Hind III site without affect- 
ing the viability of the plasmid. Frag- 
ments from a complete Hind III digest 
of 5S DNA were joined to Hind Ill- 
digested PSC101 DNA with E. coli poly- 
nucleotide ligase, and the mixture was 
used to transform tetracycline-sensitive 
E. coli. Individual tetracycline-resistant 
clones were isolated and hybrid plasmids 
purified and analyzed. As expected, the 
single repeat of 5S DNA from each clone 
was homogeneous and different in length 
from those derived from other clones 
(Fig. 4). 

We then prepared hybrid plasmids 
containing several repeat units of 5S 
DNA. To ensure that all the repeats in 
a single plasmid were adjacent in the 
genome, a partial Hind III digest of 5S 
DNA was prepared and fractionated by 
gel electrophoresis. The bands containing 
4 and 5 repeats were excised from the 
gel and the DNA eluted. After checking, 
by electron microscope measurements, 
that each preparation contained only 
DNA of the expected size, they were 
joined to PSC101 and used for transfor- 
mation. As before, hybrid plasmid DNA 
was prepared from individual clones. We 
checked to see that these plasmids had 
the anticipated structure. The plasmids 
were cut with EcoRI (5S DNA has no 
EcoRI site; while the single site in 
PSC101 is very close to the Hind III 
insertion point) , partially denatured 
with alkali, and prepared for electron 
microscopy. Each molecule showed the 
characteristic 5S repeating structure, 

Fig. 4. Electrophoresis in a 2.0% polyacryl- 
amide, 0.5% agarose composite gel of Hind III 
digests of hybrid plasmids containing single- 
repeat units of 5S DNA. From left: three indi- 
vidual plasmids, a mixture of the three, and 
total 5S DNA. 

with the appropriate number of bubbles, 
covalently joined to the plasmid (Fig. 5). 
The distribution of repeat length within 
these plasmids — and therefore within the 
chromosomal segments — was analyzed 
by electrophoresis of their Hind III 



• • ■* •• .*". — • "-V" •***.•■-**"/.*»■.-*...• " , v; •:•,•-* •><"",.-."• .-%/**-. T. :»*•'- ■*.-. v 

•-,*.* - f -i*" '-/ •. *** , «v '••"•>*. - * V* ' «r if ••**•.*»*,?* * A; % * •'. * •''".'.»;• 

"tit. • * r, 4 i*V^»,«. :• *«•-*'•.*'.'"•*■»* *,**-»,.*%'".»"* A *.»-» **»... '/*•**' *#.»"*;." *■*".*/. : -" J " .T--* ■'-'» • 
• *»#•*.»-«* **«,**» -'-■<.»» '■'**"•«»«• » « ■«' j» • »» «j ♦ * »t' , "*»j _-..**'*•«« 

" .. v . i *• *• « " , * • *« * * **»«*« « ' « ,* . « * "»*■*.. v . « • , * -.." ■ v*~»» '"* * 

" fi"i***-» -*'*'"'"« ."*'*. *"".'*%' **.** •• > * '*"»*" %■* f -**■ ' '••*•** j, *-. ' -, ' » * "*>»*"» *,'*.''*'"". »'; •' "'. ttyt* ■", ■ 

'»»"# »'***"'*- * *i" *"•'•■ » ».*."*»* f ■* ' /■*"♦ %>*" *?"-*.* »t '"*""*« >3 T " *■'/•''***»."•' ** V '. 

/?!?• !*"- ' V %* % ■*"* *.**»' •*.*'* *•*■' "*'*."•** -t" •"*.*» * «"** *« r - t *-'.*\ *.*"**"■' "'.,'j.-'. *•'* "'*'*'- ■* '" *' '*<-*'r 
! < V* ' • " '? :"' -'1 »%**> "*'? -"**,**•*•»'>'-'«** '»# ~* *!■'•* ,*•••' ■'-*•-'"' ; '»0 ; .'.** I**."- \/.*!^--~*'- : *"•*..*/,•.'** 

*'v* * » . * * * ** *. * * " «* «» 1 <■*>»«« -'^** **<» ** <*v '"*"•>, " * ■»»* * v «"<.'* *»* * ■ * '." » ■ ',; *'**'■»; !***»**••* 

«... - *. » »* . ■ , »»> .- f t '■" » - . * «* . * « * i * •.*.,*'•"«,• .- »* * » * « 1 «a. • ••. >.*. 

i vv^ ; 7 * - ; ; • ■* * - » v W f - 1>> *^ ► ~. • .v- * '• •- ;■- .?-•.- * > « ? *■• * • v-w \ -• - «- ^ I 

,■*« * *.. * ***** 4* . .»■. *■» « *••*-,., , -in,-**.' » -r * 4 . "* *•• **i' 

" -', *'* C- .**»'"' »*•» - « % V*.* ' •*"' *''*»' «*»*»•x , V *' #'»'•*»**•*■"" **•»«* « -•""*■* ,*1 » * V.'*'*'; '«*'-* 5k - M JZ "*<• 16 
■•"♦':.'. * *v >;*'."* *'C-' *■ *. ♦'**/« ' ♦•**'"" .-** iv>*< - ■• ^ .**•-•*'♦*•..* *, ,*J .♦*/• •*■**..♦: '»'. 

« ' » » ** » » *- #* *^ ** « * ^ ** * *^***.^^*»'' y y# « >. j «- ^ . ^ ■ % \ *•»***« _* ** * * ' 

. *"*** t*- *<*',«# "' ." j <*'..'* ' %*♦»'*"'»•*•'-"**•'•>"*»**"♦» "«- • • *"**■,«**»*"*•»'* ,*»'*?*]. 

'.'"••-•"•**-'*V'*' *: . . * »*"*■•* ' •- «'5tr* «,«* > **« ,. -v v j ' '• r , ,»'*.<»* .' **. *' • '■•»?- 

- • ..-*•;**«< ** ' "«»'-*■'♦ *',,!" | * v -'. ,""*"•'•'.'••' i - >* • *•*>*.•*"* • i . - '. *"*.;' V. 

% . *s.* v y*,«* ~ ; *,' « ; '- ; ■- » '* % * j,*?? *.*^-' * V *]'■'*.♦* ' 7 • * -'■♦* ? -*.. •'*.'•'-*■' o .y. : -> *%»V-"** r> v "*, 

- ».' ** V* •*" *-*.>• e '•'. - * - ■■» .- J* * C" "'*.»» * ' -" * ' * * •- . ■** i "' *' r " •''■•♦ , 

;.;.";% j i- ,, **.V«'>;-' 4 *•/»".•.»« t*.^*'-c *•'•** ^'* ^." w/VW^. ,*r* **'/- •'•"r *•'-«> *r *' 

Fig. 5. A partially denatured molecule of a hybrid plasmid DNA in covalent linkage with 
5-repeat units of 5S DNA. The circular molecule has been linearized by EcoRI cleavage. 
The PSC101 portion is at the top, as shown, and contains three denatured regions. The 5S 
DNA repeats are seen at the bottom. 



digests. Each plasmid yielded a char- 
acteristic pattern with several different 
5S DNA repeat units (Fig. 6). Therefore, 
adjacent repeat units are not in general 
identical; and sudden correction mecha- 
nisms (such as the hypothetical master- 
slave mechanism) cannot be operating 
in this DNA. 

A "gradual" mechanism of tandem 
evolution which has proceeded by un- 
equal crossing-over has several attractive 
features: (1) It does not require the 
postulation of any unknown mechanism. 

(2) It can spread or eliminate variants. 

(3) The presence of multiple repeats 
(both full repeats and internal, 15-mer 

Fig. 6. Electrophoresis (as in Fig. 4) of Hind III digests of hybrid plasmids containing 
4 and 5 repeat units of 5S DNA. From left: a Hind III digest of SV40 DNA, total 5S DNA, 
two different 4-repeat plasmids, two different 5-repeat plasmids, a Hae III digest of SV40 


repeats) may increase the crossover fre- gene. These will be used as hybridization 

quency compared to nonreiterated DNA. probes of transcripts made from Xenopus 

(4) Any crossover may be unequal not laevis 5S DNA in vitro. It is hoped that 

only with respect to full repeat units but in this way a sensitive and specific assay 

also with respect to internal repeats. In for authentic transcription in vitro may 

fact, recombination with out-of-phase be developed, 
pairing of the 15-mers would generate 

exactly the type of repeat length hetero- Xenopus laevis-Xenopus mulleri 

geneity we observe. Hybrid Animals 

D. D. Brown and E. Jordan 

Transcription of 5S DNA __ „_. . , v 

In Year Book 72 (p. 11) and Year 

Scott Emmons Book 73 ( p 9^ we described the use of 

To understand the control of gene hybrid animals to "introduce hetero- 

expression in eukaryotes it will be neces- geneity" into the ribosomal DNA locus, 

sary to reconstruct their transcription in When Fi animals were backcrossed with 

vitro. We are attempting to find out in wild-type Xenopus of either species, only 

this way how the 5S genes of Xenopus a very few wild-type segregants were 

are expressed. found. This evidence supported the no- 

The 5S genes are transcribed in vivo tion that extensive homologue exchange 

by some form of RNA polymerase III must have occurred in the F 1 germ cells. 

(Weinmann and Roeder, Proceedings of These experiments have not been con- 

the National Academy of Sciences, 71, tinued for three reasons: (1) An inde- 

1790-1794, 1974), and the properties of pendent method for demonstrating cross- 

this enzyme purified from Xenopus ing-over between the laevis and mulleri 

have been studied extensively (Sklar, rDNA could not be devised and nondis- 

Schwartz, and Roeder, ibid, 72, 348-352, junction could not be ruled out. (2) 

1975). However, when presented with Extensive crossing-over between rDNA 

isolated 5S DNA, this enzyme does not in the two species might not mimic the 

make 5S RNA (Roeder, personal com- normal rate of crossing-over between 

munication). The problem therefore is homologues of one species. (3) New 

to find a different form of either the methods for analyzing crossing-over 

enzyme or the template which endows within a single species became available, 

the system in vitro with the specificity involving restriction enzyme digestion of 

existing in vivo. In beginning our search, purified ribosomal DNA, a technique 

we are developing an assay for correct that gives a kind of "fingerprint" of 

initiation of transcription at the begin- rDNA in a single nucleolar organizer 

ning of the 5S gene. Our scheme involves locus. Results of these experiments are 

the use of restriction enzyme fragments described for X. laevis ribosomal DNA 

of 5S DNA from Xenopus mulleri which elsewhere by P. K. Wellauer et al. (this 

carry the first few nucleotides of the 5S Report). 



Xenopus laevis RIBOSOMAL DNA 

P. K. Wellauer, R. H. Reeder, I. B. Dawid, and D. D. Brown 

Tandem ribosomal genes are separated species of Xenopus, suggesting that the 

from each other by spacers of unknown function of spacers does not rely on a 

function. Most of the sequence of these specific nucleotide sequence. Yet the mul- 

spacers evolves at a high rate between tiple spacers in a tandem array of genes 



and spacers resemble each other closely; 
they have evolved together. In the ex- 
periments reported here, we have sought 
to obtain detailed information about the 
organization of spacer sequences and to 
distinguish between two general models 
which would account for this tandem 
evolution in the ribosomal DNA of 
Xenopus. Sudden correction mechanisms 
predict that a mutation in one repeat 
will be spread to adjacent repeats in a 
single event. Examples of this are the 
"master-slave" mechanism and the ex- 
pansion and contraction of multiple re- 
peats. One likely "sudden" mechanism 
is by "rolling circle" replication in which 
one repeat is replicated continually, giv- 
ing rise to many identical tandem repeats. 
Gradual mechanisms such as multiple 
unequal cross-overs between homologues 
and/or sister chromatids predict a grad- 
ual scrambling of new sequences that 
arose by mutations, fixing some in the 
population and eliminating others. We 
have attempted to distinguish between 
these two general mechanisms by analyz- 
ing the newly discovered length hetero- 
geneity within the multiple repeats of 
ribosomal ,DNA. We have asked: (1) 
What is the molecular basis of this 
length heterogeneity? (2) How is it ar- 
ranged within a single nucleolar orga- 
nizer? and (3) Is the pattern of length 
heterogeneity inherited "intact" or is it 
"scrambled" in a single generation? Sud- 

den correction mechanisms predict tan- 
dem homogeneity. Slow mechanisms per- 
mit scrambling of heterogeneity between 
adjacent repeats. 

The Organization of Sequences 

within the Spacer of Xenopus laevis 

Ribosomal DNA 

Each repeating unit of ribosomal DNA 
(rDNA) consists of a gene region that 
is transcribed into the 40S ribosomal 
RNA precursor plus a nontranscribed 
spacer region (Fig. 7). While all 40S 
RNA regions in X. laevis are identical 
at the resolution provided by our present 
technology, the spacers vary in length 
{Year Book 73, p. 35). To study the 
basis of these length variations we took 
advantage of the development by Mor- 
row et al. (Proceedings of the National 
Academy of Sciences, 71, 1743, 1974) of a 
method of linking fragments of rDNA to 
a bacterial plasmid and then cloning and 
replicating the hybrid plasmid in E. coli. 
The restriction endonuclease EcoRI cuts 
rDNA at two sites, yielding two frag- 
ments of DNA from each repeating unit. 
The smaller of these is homogeneous and 
comprises most of the transcribed region. 
The larger of the two contains all of the 
nontranscribed spacer. Since the spacer 
regions are heterogeneous in length, this 
fragment also varies in size. Four cloned 
spacer-containing fragments of different 
lengths were used in this study. Two of 










3.0 X I0 6 3.8-7.5 X I0 6 

EcoRI EcoRI EcoRI 

Fig. 7. Model of one repeating unit of Xenopus laevis ribosomal DNA. The repeat begins 
at the 5'-end of the 40S transcription unit. The 40S precursor region contains the 28S and 18S 
genes plus two transcribed spacers (see Year Book 73, 45). The length variability is restricted 
to the nontranscribed spacer. The restriction endonuclease EcoRI cuts each rDNA repeat at 
two sites as indicated by the arrows. 



these with molecular weights of 3.9 and 
4.2 X 10 6 had been generated by Morrow 
and his colleagues and two more (mo- 
lecular weights of 5.4 and 6.6 X 10 6 ) 
were prepared by similar methods. 

These cloned homogeneous rDNA 
fragments were analyzed by homoduplex 
and heteroduplex mapping in the electron 
microscope. The main conclusion of this 
work is that the spacer region is com- 
posed to a large extent of internally 
repetitious units. These subrepeats are 
limited to two regions of the spacer 
(designated A and C in Fig. 9). We do 
not know the size of the subrepeats, but 
they are probably less than 50 nucleo- 

The evidence for internal repetition 
within the nontranscribed spacer is based 
on the following observations. If two 
cloned rDNA molecules of different 
lengths had identical, unique sequences 
except for an extra DNA piece which 
accounted for their difference, a hetero- 
duplex between the two molecules should 
always result in a double-stranded region 
corresponding in length to the shorter 
strand and a single-stranded loop at a 
fixed position, the point where additional 
sequences are present in the longer 
strand. However, heteroduplexes between 
rDNA fragments showed multiple loops 
and the position of these loops varied, as 
illustrated in Fig. 8. This situation can 
be explained if the region within which 
loops occur is tandemly repetitious. Loops 
with variable positions have been ob- 

served by Busse and Baldwin {Journal 
of Molecular Biology, 65, 401, 1972) in 
heteroduplex molecules involving bac- 
teriophage DNAs with duplications. 
Measurements of heteroduplexes between 
cloned rDNA molecules defined two dis- 
tinct regions (A and C) of repetitive 
DNA (Figures 8 and 9), separated by a 
segment in which no repetitiveness could 
be detected (region B). In addition, the 
spacer segment adjacent to the 5'-end of 
the transcription unit (region X) is de- 
void of loops in heteroduplex molecules 
and therefore is presumed to lack se- 
quence repetition. From its location in 
the repeat, region X probably contains 
various control sequences which are in- 
volved in RNA transcription. 

The length heterogeneity between 
repeats is confined to regions A and C 
of the spacer. Thus, long rDNA repeats 
have more copies of subrepeats than do 
short rDNA repeating units. Confirma- 
tion of this idea has been obtained from 
thermal denaturation profiles of cloned 
rDNA fragments of different lengths. 
Part of the hyperchromicity occurred 
within one very sharp transition. The 
fraction of the total hyperchromicity 
that occurred in this transition increased 
from the shortest to the longest rDNA 
fragment. This hypersharp melting tran- 
sition represents primarily the denatura- 
tion of region A. Such hypersharp melt- 
ing curves are characteristic of simple 
sequence DNAs like satellite DNAs. 











Fig. 8. Model for heteroduplex molecules between rDNA fragments of 3.9 and 5.4 X 10 8 
daltons. Since regions A and C are composed of small subrepeats, loops formed in these 
regions can occur at various positions by association of the strands in different registers. 





NTSp 1.8 

TSp A B C X 28S 

h h ^ 1 \ — i 


->+•« p+4 M«« »H » M W 

0.55 0.53 0.44 027 0.43 




TSp A 

H 1 I- 


B C X 28S 

1 -\ 1 


I8S TSp Y 

1 »— I- 



B C X 28S 

1 H 1 

2.0 [1.7] 





TSp Y 
H •— I- 

X 28S 

■\ 1 


Fig. 9. Models for the four cloned rDNA fragments which were studied. The molecular 
weights of the fragments are: CD30, 3.9 X 10 6 ; CD42, 4.2 x 10 e ; Xlr4, 5.4 X 10 6 ; Xlr5, 6.6 
X 10 6 . The size of each region is given in CD30 and again in the larger fragments whenever a 
change in the size of a particular region was found. TSp stands for transcribed spacer, NTSp 
for nontranscribed spacer. 

The Arrangement of Length 

Heterogeneity in Amplified and 

Chromosomal rDNA 

In this section we address several 
questions which arise from the finding 
that X. laevis rDNA contains spacers 
of various lengths. We ask whether 
length heterogeneity occurs in single in- 
dividuals and within one nucleolar orga- 
nizer, and if this occurs, whether repeat- 
ing units of different lengths are arranged 
in blocks of like units or whether they 
may be scrambled. We further compare 
the distribution and arrangement of dif- 
ferent size classes of spacers in chromo- 
somal and amplified rDNA. 

Figure 10 shows representative electro- 
phoretic patterns of rDNAs which had 
been purified from individual frogs and 
digested with EcoRI restriction endo- 
nuclease. Each pair of gels compares 
amplified and chromosomal rDNA from 
a single frog; each pair displays two 
patterns of about the same complexity. 

There are a few generalizations that can 
be drawn from these experiments. 

First, bands that are visualized in 
amplified rDNA in most cases correspond 
to bands in chromosomal rDNA, indicat- 
ing that no additional heterogeneity is 
introduced during amplification. The 
only exception to this rule found so far 
is shown in gel b of Fig. 10. 

Second, the gel analysis demonstrates 
that there may be strong selection for 
certain repeats during amplification. For 
instance, the chromosomal rDNA in gel 
a (Fig. 10) shows two strong bands of 
spacer containing rDNA fragments with 
molecular weights of about 4.0 and 5.0 
X 10 6 plus a faint band at about 4.5 X 
10 6 . The amplified rDNA derived from 
this chromosome has only one strong 
band at 5.0 X 10 6 plus two faint bands 
at 4.0 and 4.5 X 10 6 . In one chromosome 
the most abundant class of repeating 
units may become amplified, while in 
another chromosome a rare class of re- 
peats will undergo amplification. This 



observation may also explain the electro- 
phoretic patterns shown in gel 6, since 
one could argue that the additional band 
which is observed in amplified rDNA 
from this frog may be present in chromo- 
somal rDNA as well but at too low an 
intensity to be seen. 

Third, although we do not know the 
reasons for this selective amplification, 
we can demonstrate that it must be 
genetically determined. Gels c and d 
(Fig. 10) show that sibling frogs which 
inherited the same chromosome amplify 
their rDNA in the same way. 

These observations raise the question 
whether the intramolecular arrangement 
of length heterogeneity is the same in 
amplified and in chromosomal rDNA. 
We approached this problem in the fol- 
lowing way. Long single strands from 
two preparations each of amplified and 
chromosomal rDNA were hybridized 
with an excess of the cloned 3.9 X 10 6 

a b 

spacer-containing rDNA fragment, and 
the molecules were analyzed in the elec- 
tron microscope. Representative micro- 
graphs and tracings of two such hetero- 
duplex molecules are presented in Figs. 
11 and 12. The amplified rDNA molecule 
(Fig. 11) displays three and the chromo- 
somal rDNA molecule (Fig. 12), two 
heteroduplex regions which are separated 
by segments of single-stranded DNA. 
One heteroduplex region plus one single- 
stranded region add up to one rDNA 
repeating unit. The three heteroduplexes 
in the amplified rDNA were identical in 
size, whereas the heteroduplexes in the 
chromosomal rDNA differed. Measure- 
ments of a large number of strands from 
the two preparations of amplified rDNA 
demonstrate that, within the limits of 
this technique, between 87% and 98% 
of nearest neighbor repeats are identical 
in size within one particular strand of 
DNA. On the other hand, only between 

d e 






Fig. 10. Comparison of EcoRI digests of amplified (A) and chromosomal (C) rDNA from 
individual frogs which contain a single nucleolar organizer (heterozygous frogs carrying the 
anucleolate mutation). The rDNAs were isolated from four separate frogs, digested with 
EcoRI restriction endonuclease, and the fragments electrophoresed on 1% agarose gels. The 
rDNAs from each frog are grouped together. Gels a and b show the patterns of two sibling 
frogs which inherited different chromosomes; gels c and d show two siblings which inherited 
the same chromosome; gel e contains Hind III digested phage X DNA. Molecular weights 
of the fragments are multiplied by 10" 6 . 



f >U. 

h* v - 

Fig. 11. Electron micrograph and tracing of a molecule of amplified rDNA. This molecule 
displays three full repeating units of identical size. Single strands of rDNA were hybridized 
with a threefold excess of the 3.9 X 10 8 spacer-containing fragment and were examined by 
electron microscopy. 


.-• Y 

.'.V • 

•V ; 

Fig. 12. Electron micrograph and tracing of a molecule of chromosomal rDNA. The two 
repeating units in this molecule are of different lengths. The molecule was analyzed as 
described in the legend to Fig. 11. 

30% and 50% of the nearest neighbor 
repeating units of chromosomal rDNA 
are identical in size. In both samples of 
chromosomal rDNA, 30% identical near- 
est neighbor repeats would be present if 
repeats of different size classes were 
arranged at random in these DNAs. 
Therefore, in one of the chromosomal 
rDNAs analyzed, the arrangement of 
length heterogeneity is compatible with 
random expectation. 

The main conclusion from these ex- 
periments is that the intramolecular ar- 
rangement of length heterogeneity is 
fundamentally different in amplified and 
in chromosomal rDNA. The observation 
that most if not all repeating units within 
one molecule of amplified rDNA are 
homogeneous in size is compatible with 
a rolling circle mechanism for amplifi- 
cation. The data also support a model 
in which predominantly single repeating 



units are copied or excised from chro- 
mosomal rDNA during the initial ampli- 
fication event. 

The Inheritance of rDNA Spacer 
Length Heterogeneity 

We have begun a series of experiments 
to learn how rDNA spacer patterns are 
inherited. In particular, we are interested 
in observing how stable the spacer pat- 
terns are and what types of changes 
might be detectable from one generation 
of animals to the next. Such experiments 
might give clues to the mechanism by 
which homogeneous and conserved gene 
regions are maintained interspersed with 
more rapidly evolving and variable 
length spacer regions. 

So far about 50 progeny from three 
separate matings have been analyzed. 
The restriction enzyme patterns of puri- 
fied chromosomal rDNA of the parents 
and selected offspring from one mating 
are shown in Fig. 13. In this cross a 
female heterozygous for the anucleolate 
mutation (1~1) was mated with a wild- 
type male (11). To simplify analysis, 
only heterozygous offspring (1~1) were 
studied. The rDNA fragments of all but 
two of the offspring fall into two groups 
of different fragment patterns (Fig. 13c, 
d). The sum of these two patterns gives 
the pattern of the male parent's rDNA 
and therefore represents the normal 
segregation of the two male chromo- 
somes. Since only 1~1 offspring were 
analyzed, the female's single rDNA- 
containing chromosome was excluded 
from the selected progeny. 

Two of the offspring had spacer pat- 
terns which were related to the two 
major classes but differed significantly, 
One of these patterns differed from the 
pattern shown in Fig. 13c, by having 
its 5.1 X 10 6 class of spacer fragments 
doubled in quantity compared to the 
other spacer classes. In the other unusual 
case the pattern was similar to that 
shown in Figure 7d, but it contained an 
extra band at 4.3 X 10 6 . At present we 
do not know how either of these unusual 

fig. 13. EcoRI fragment patterns of chromo- 
somal rDNA from parents and offspring of a 
single mating. A female heterozygous for the 
anucleolate mutation (l'l) was mated with a 
wild-type male (11) and offspring bearing the 
mutant chromosome were analyzed. Chromo- 
somal rDNA was isolated from each frog, di- 
gested with EcoRI, and the digest electro- 
phoresed on a 1% agarose gel. After staining 
with ethidium bromide, the gel was photo- 
graphed and the photograph traced in a densi- 
tometer. Densitometer traces are shown of 
rDNA from (a) the female parent; (b) the 
male parent; (c) one class of offspring; (d) the 
other major class of offspring. The spacer- 
containing bands are identified by MW X 10"°. 

patterns arose. Unequal crossing over 
(either homologue or sister chromatid 
exchange) seems unlikely, since it would 
require the spacer size classes to be 
arranged in homogeneous blocks in the 
chromosome. The heteroduplex analysis 
described earlier in this report suggests 



that such homogeneous blocks do not 

The general conclusion at this stage is 
that the spacer patterns are usually 
transmitted from one generation to the 
next without detectable alteration. In 
order to understand the mechanism that 
gives rise to occasional changes in rDNA 
restriction pattern, we are now analyzing 
additional crosses. 


These experiments provide partial or 
complete answers to the three questions 
posed in the introduction. (1) Length 
heterogeneity exists only within specific 
regions of the nontranscribed spacer re- 
gion. The greatest variation in length 
lies within the spacer region adjacent to 
the 3'-end of the transcription unit. 
Those regions which are heterogeneous 
in length consist of simple, highly re- 
iterated sequences. Fewer or greater 
numbers of these sequences account for 
the length variation between spacers. 
(2) Single nucleolar organizers having 
several hundred tandem repeats can con- 
sist of as many as seven distinguishable 
repeat lengths. Analysis of adjacent re- 
peats by heteroduplex mapping shows 

extensive scrambling of this hetero- 
geneity. Thus, in a high proportion of 
the repeats, the length of one repeat is 
different from that of its neighbor. In 
contrast, the multiple repeats along a 
single molecule of amplified rDNA are 
homogeneous in length. This is in good 
agreement with the observations of 
others that amplification occurs by a 
rolling circle mechanism. (3) The char- 
acteristic repeat length heterogeneity 
within a nucleolar organizer is generally 
inherited without change by the progeny. 
This was found to be the case in all but 
two animals from three matings. 

We conclude that the explanation for 
length heterogeneity in rDNA is the 
same as that for 5S DNA: it is caused by 
the presence of variable amounts of 
reiterated simple sequences in the non- 
transcribed spacer. This region of vari- 
able length, like the one in spacer regions 
of 5S DNA, is close to if not adjacent to 
the 3'-end of the transcription unit. 
Adjacent repeats of rDNA like those in 
5S DNA (D. Carroll, this Report) are 
frequently of different lengths. This 
rules out the "sudden" correction mecha- 
nisms in their extreme form and supports 
the crossing-over model as the means 
by which tandem repeating sequences 
evolve together. 



R. H. Reeder, T. Higashinakagawa, and S. Biroc 
with the assistance of E. Hogan 

In Year Book 73, p. 28, Higashinaka- 
gawa reported the isolation of highly 
purified ribosomal gene chromatin from 
Xenopus oocytes that are active in ribo- 
somal RNA synthesis. This year he re- 
ports his first efforts at characterizing 
the proteins associated with the active 
ribosomal genes. He has found a signifi- 
cant amount of histone in the ribosomal 
gene chromatin, and at least some of 
this histone appears to be bound to the 
actively transcribed gene regions. 

Eventually we hope not only to pro- 
duce a complete catalogue but also to 
map the locations of all the proteins 
associated with these genes, both when 
they are transcribed actively and when 
they are inactive. 

Parallel experiments have been done 
with the ribosomal gene chromatin from 
Xenopus cultured cells. It was not possi- 
ble to physically isolate the ribosomal 
gene chromatin. However, experiments 
using limited DNase digestion of isolated 



nuclei have shown that the ribosomal 
genes are partially protected from DNase 
in a manner similar to the bulk of the 
chromatin, thus providing indirect evi- 
dence that the active ribosomal genes 
in cultured cells are also associated with 

In a separate line of research, Biroc 
has examined the accessibility of histone 
tyrosines in chromatin to reaction with 
radioactive iodine. The reactivity to 
iodine changes in a specific manner as 
the chromatin structure is progressively 
taken apart. These studies lay the 
groundwork for detection of possible 
conformational differences of histones on 
active versus inactive chromatin. 

The Association of Ribosomal Genes 
with Histones in Cultured Cells 

R. H. Reeder 

This year has seen an explosion in the 
literature on the structure of chromatin 
of eukaryotic organisms. The finding, in 
several laboratories, that chromatin 
possesses a repeating subunit structure 
has been especially compelling. Histones 
have been shown to be associated with 
the DNA chain, forming a structure that, 
in the electron microscope, resembles 
beads on a chain. Each bead consists of 
about 200 base pairs of DNA plus a 
specific cluster of histone molecules. 
When chromatin is digested briefly with 
a nonspecific DNase such as micrococcal 
nuclease, the DNA in the bead is rela- 
tively resistant to attack, while the DNA 
connecting the beads is relatively open. 
Thus, brief DNase digestion results -in 
cleaving most of the DNA into 200 base- 
pair fragments or oligomers. These ob- 
servations suggest an indirect method of 
learning whether or not specific genes 
are associated with histones. One can 
treat unfractionated chromatin (or iso- 
lated nuclei) with DNase, separate the 
resultant DNA fragments by gel electro- 
phoresis, and locate fragments contain- 
ing a specific gene by molecular hybridi- 
zation with a radioactive RNA specific 

for that gene. If the gene in question is 
cut into 200 base-pair pieces or higher 
order oligomers, that can be taken as 
good evidence that it is associated with 

We have used this protocol to study 
the genes for 18S and 28S ribosomal 
RNA in Xenopus cultured cells. Nuclei 
labeled- with 3 H-thymidine were isolated 
from cells that had been seeded in roller 
bottles four days before and were grow- 
ing maximally. This particular cell line 
is not contact inhibited and grows until 
it exhausts the medium. Aliquots of the 
nuclei were digested with various 
amounts of micrococcal nuclease, and 
the DNA fragments were purified and 
electrophoresed on acrylamide gels cross- 
linked with ethylene diacrylate. Figure 
14 shows the distribution of the 3 H- 
labeled DNA fragments after three dif- 
ferent degrees of DNase digestion. As 
reported by others, the bulk of the DNA 
is cut into an oligomeric series of frag- 
ments. To locate the ribosomal genes in 
such gels, each gel was sliced, each slice 
was hydrolyzed with alkali to liberate 
the DNA, and the DNA from each slice 
was hybridized in solution with 32 P- 
labeled ribosomal RNA. After hybridi- 
zation the hybrids were trimmed with 
RNase and caught on a nitrocellulose 
filter to be counted. The distribution of 
the 32 P label in Fig. 14 shows that the 
ribosomal genes were cut into the same 
fragment sizes as was the bulk DNA, 
strongly implying that they are associ- 
ated with histones in a manner similar 
to that of the bulk chromatin. Other 
experiments (not shown) indicate that 
the spacer region of the ribosomal DNA 
is also associated with histones. One 
problem in interpreting these results is 
that we have no direct measure of what 
fraction of the ribosomal genes are ac- 
tively transcribed even in a rapidly 
growing cell. One might argue that the 
active genes are only a fraction of the 
approximately 900 ribosomal genes in 
the cell and that they are preferentially 
destroyed very early during DNase di- 






i 600 

x 400 







|_ d ^V°e \ 




- 1000 

10 20 30 

Slice Number 


Fig. 14. Hybridization of cultured cell nuclear 
DNA fragments with ribosomal RNA. Cultured 
Xenopus cells were grown for 4 days in roller 
bottles and labeled overnight with 3 H thymi- 
dine, and nuclei were isolated. Aliquots were 
digested with increasing amounts of micrococcal 
nuclease in 0.01 M Tris pH 8, 0.01 M NaCl, 5 
ml MgCl 2 for 10 min at 37°C. The DNA frag- 
ments were purified and electrophoresed on 6% 
acrylamide tube gels cross-linked with ethylene 
diacrylate. The gels were sliced, each slice 
hydrolyzed with alkali, and the DNA in each 
slice hybridized with 32 P rRNA. The hybrids 
were trimmed with RNase and caught on nitro- 
cellulose niters in 3 M NaCl. (a) 16% digestion 
of the bulk DNA, (b) 33% digestion, (c) 41% 
digestion. The major peak at about slice no. 
30 has a size of 135-185 base pairs depending 
upon the strength of DNase digestion. 

gestion. If that were true, then Fig. 
14 would be telling us only about the 
remaining inactive genes. This interpre- 
tation seems unlikely, since control ex- 
periments have shown that the ribosomal 

genes are progressively destroyed by 
DNase, as a "smooth function" of diges- 
tion time. We cannot detect the rapid 
early loss of a fraction of these genes. 
Furthermore, nuclei from young active 
cells and from old cells that have ex- 
hausted the medium and have stopped 
growing both show the same linear 
kinetics of ribosomal gene digestion by 

Another possible artifact could have 
arisen if histones were able to rearrange 
during the isolation and digestion of the 
nuclei. This artifact has been partially 
ruled out by an experiment in which a 
trace amount of radioactive deproteinized 
DNA was mixed with a large excess of 
unlabeled nuclei and the mixture digested 
with DNase. While the unlabeled DNA 
in the nuclei was digested into the usual 
oligomeric fragments, the deproteinized 
radioactive DNA was rapidly destroyed 
without being cut into any discrete size 
classes, suggesting that rearrangement of 
histones is not a serious artifact in these 

In summary, therefore, we feel that 
the most likely interpretation of these 
experiments is that histones are associ- 
ated with active ribosomal genes. 

Analysis of the Proteins of 
Isolated Ribosomal Gene Chromatin 

T. Higashinakagawa 

Ribosomal gene chromatin has been 
isolated in a highly purified state from 
the nucleoli of Xenopus laevis oocytes 
(Year Book 73, p. 28). By this tech- 
nique we are now able to provide chro- 
matin of 75% -80% purity routinely — 
and occasionally even of 95%-100% 
purity. Since the nucleolus is an indicator 
of active ribosomal RNA synthesis, 
analysis of this chromatin should give 
us information about the proteins on 
active genes, just as studies with bulk 
DNA chromatin should tell about the 
proteins on inactive genes. We have 
begun to analyze the protein moiety of 
rDNA chromatin with special attention 
to the presence or absence of histones. 



Ribosomal gene chromatin was dis- 
solved in SDS and electrophoresed on an 
SDS-polyacrylamide slab gel together 
with proteins of various origins for the 
purpose of comparison (Fig. 15). As 
shown in Fig. 16, this particular prepara- 

MW 12 3 4 5 6 





. Mfe 






Fig. 15. SDS-polyacrylamide gel electro- 
phoresis pattern of ribosomal gene chromatin 
proteins. Ribosomal gene chromatin and pro- 
teins from other sources were dissolved by heat- 
ing in SDS and /3-mercaptoethanol and electro- 
phoresed on a slab gel of 15% polyacrylamide- 
0.1% SDS. (1) Molecular weight markers, (2) 
total nucleolar protein, (3) ribosomal protein, 
(4) ribosomal gene chromatin protein, (5) 
Xenopus bulk DNA chromatin protein, (6) 
Xenopus acid extracted histone. An arrow in 
slot (4) points to the possible RNA polymerase 
I subunit. 

tion happened to possess no detectable 
amount of bulk chromosomal DNA when 
examined in an analytical CsCl density 
gradient; it was therefore almost 100% 
pure ribosomal gene chromatin. The 
ribosomal gene chromatin clearly con- 
tains histones F 2al , F 2a2 , F 2b and F 3 (Fig. 
14, slot 4). There is no discrete band 
which corresponds to histone F 1 in this 
gel. Another band of interest is visible in 
the high molecular weight region of the 
gel (marked by the arrow). This rather 
strong band has almost the same mobility 
as that of the E. coli RNA polymerase 
PP' subunits, and hence may possibly 
represent a subunit of the RNA poly- 
merase I molecule. Almost all other pro- 
tein bands coincide with ribosomal pro- 
tein or total nucleolar protein and are 
present in such large amounts in the 
nucleolus that they are unlikely to be 
specifically bound to the ribosomal DNA. 
At this resolution we can detect no other 
rDNA-specific protein that is present in 
the same abundance as histones. Our 
techniques are not yet sufficiently sensi- 
tive to detect proteins which may be 
present in only a few copies per rDNA 
repeating unit. 

We wished to measure the histone-to- 
DNA ratio in rDNA chromatin. Since 
the yield of this chromatin is low 
(0.1-0.2 fAg of rDNA from one frog), 
conventional analytical procedures are 
difficult to apply. However, we have 
developed a procedure in which electro- 
phoresis is carried out in 2 mm diameter 
gel columns, and in this way 0.5-1.0 fxg 
of histones can be visualized and quanti- 
tated by staining. Figure 17 shows the 
electrophoretogram of the ribosomal gene 
chromatin protein carried out in such a 
small gel column. By comparison with a 
gel like the one shown in Fig. 15, each 
histone component can be assigned on 
this gel as indicated in the figure. The 
amount of histone was estimated from 
the area under the stained peaks while 
the amount of DNA as well as the per- 
centage of rDNA (purity) was measured 
in an analytical CsCl density gradient 



bulk DNA 

Fig. 16. Analytical CsCl density gradient profile of DNA from the ribosomal gene chromatin 
analyzed in Fig. 15. Virtually no bulk DNA is detectable, showing that this is almost 100% 
pure ribosomal gene chromatin. 


f 3 » f 2a2 



Fig. 17. SDS-polyacrylamide gel electrophoresis pattern of ribosomal gene chromatin protein 
carried out in a 2 mm diameter gel column. The conditions of electrophoresis were the 
same as in Fig. 15. This gel corresponds to experiment 3 in Table 2. "X" indicates the location 
of the possible RNA polymerase I subunit. 

like that shown in Fig. 16. The entire 
measurement was calibrated by con- 
structing a standard curve using bulk 
DNA chromatin prepared from Xenopus 
cultured cells under exactly the same 
conditions. The results of these experi- 
ments are summarized in Table 2. 

The amount of F 2a i is smaller than is 
found in bulk Xenopus chromatin. This 

also seems to be the case with the sum 
of F 2a2 , F 2b , and F 3 . Due to some uncer- 
tainty in the overall recovery of histones, 
we regard this measurement as a mini- 
mum estimate of the amount of histone 
on the ribosomal DNA. Obviously, these 
results corroborate the finding shown in 
Fig. 15. 

The presence of histones in active 




TABLE 2. Composition of Ribosomal DNA Chromatin 

Bulk DNA rDNA Purity 

Expt. No. (fig) (fig) (%) 

Relative Histone 
Abundance* on rDNA 



-f" ^2b + F3 



















*Both bulk DNA chromatin containing a known amount of DNA (determined by Model 
E analysis) and ribosomal DNA chromatin were electrophoresed under identical conditions, 
and the area under their histone bands determined. Relative histone abundance is expressed 
as the ratio of histone areas in rDNA chromatin (after subtracting the area contributed by 
contaminating bulk DNA chromatin) to the histone area in bulk chromatin containing an 
equivalent amount of DNA. 

ribosomal DNA chromatin could be in- 
terpreted in several ways. For example, 
it is possible that the histones do not 
belong there and were accidentally 
picked up during isolation. Alternatively, 
they may belong on the rDNA but are 
only located on the nontranscribed spacer 
region, not on the transcribed gene re- 
gion. A third, and most interesting, possi- 
bility is that they are also associated 
with the active gene regions. To explore 
these possibilities we have studied the 
action of micrococcal nuclease on iso- 
lated nucleoli. 

If histones are associated with rDNA 
within the nucleolus, we should be able 
to digest the nucleoli and isolate pro- 
tected fragments similar to the protected 
fragments we described from cultured 
cell nuclei. By hybridization of these 
fragments with ribosomal RNA we 
should be able to determine if they are 
derived from the gene region, the spacer 
region, or both. Accordingly, oocyte 
nucleoli were isolated, divided into three 
aliquots, and digested with increasing 
amounts of micrococcal nuclease. After 
deproteinization, each aliquot of DNA 
fragments was labeled with 32 P, using a 
DNA polymerase exchange reaction. 
Control experiments have shown that 
this labeling method does not alter the 
electrophoretic mobility of the DNA. 
The labeled fragments were then electro- 

phoresed on a 6% polyacrylamide slab 
gel and located by autoradiography. 
Micrococcal nuclease fragments from 
whole nuclei and restriction enzyme frag- 
ments from SV 40 viral DNA were also 
labeled and run under the same condi- 
tions for comparison. The results are 
shown in Fig. 18. 

Digestion of nucleoli with small 
amounts of nuclease (lane 1) results in 
the appearance of DNA bands with sizes 
of about 135, 280, and 400 base pairs, 
plus some higher molecular weight ma- 
terial. These bands are similar in size 
to those derived from whole nuclei. We 
take their appearance as good evidence 
that histones are bound to the rDNA in 
a specific manner and are not just loosely 
associated with nucleoli due to an isola- 
tion artifact. In other experiments we 
have observed that the size of the 
nuclease-derived bands changes with in- 
creasing digestion; we assume this ac- 
counts for the size difference seen in Fig. 
18 between the nuclear and nucleolar 
fragments. With longer digestion (lanes 
2 and 3) the larger bands disappear and 
a series of smaller bands appear. Small 
fragments similar to these have been 
interpreted by other investigators as 
resulting from nuclease attack within 
the histone-DNA complex and are addi- 
tional evidence that histones are associ- 
ated with the rDNA. 





II) 12) 

\ V — — 

Fig. 18. Electrophoresis of nucleolar DNA 
fragments after partial nuclease digestion. Iso- 
lated nucleoli (72% rDNA purity) were di- 
gested with increasing amounts of micrococcal 
nuclease. The DNA fragments were depro- 
teinized and labeled with 32 P, using E. coli DNA 
polymerase, and a portion of each was run on 
a 6% acrylamide slab gel cross-linked with bis 
acrylamide. The gel was then dried down and 
autoradiographed. SV40 viral DNA digested 
with restriction enzyme Hae III and DNA 
fragments from whole nuclei digested with 
micrococcal nuclease were labeled and run at 
the same time for comparison. DNA sizes are 
given in base pairs. The sizes of the 16-35 
base-pair fragments of SV40 DNA are only 

To determine if the ribosomal DNA 
fragments were coming from the gene 
or spacer region, the rest of each sample 
was electrophoresed in tube gels of 6% 

acrylamide cross-linked with ethylene 
diacrylate. The gels were then sliced and 
the DNA from each slice was hybridized 
to saturation with 3 H-rRNA. The results 
are shown in Fig. 19. After low nuclease 
digestion (Fig. 19a) the 3 H-rRNA hy- 
bridizes with all size classes of 32 P- 
labeled fragments. Resolution of the 
various bands is not as good in diacrylate 
cross-linked gels, but the hybridization 
clearly follows the major peaks. From 

3 H 32p 








T A ' 


ft i 

d 9 

1 q 




i i i 




- 9 
\a 9. 



1 i ^ lcaoooo °o 





/ b 9 
pd q 


3 H 32 p 

10 20 30 

Slice Number 


Fig. 19. Hybridization of nucleolar DNA 
fragments with ribosomal RNA. The remainder 
of each nucleolar DNA sample shown in Fig. 18 
was electrophoresed on a 6% acrylamide tube 
gel as described in Fig. 14 and hybridized with 
3 H-labeled ribosomal RNA. Panels A, B, and 
C correspond to lanes 1, 2, and 3, respectively, 
of Fig. 17. They show the results in increasing 
degrees of nuclease digestion. 



the known specific activity of the 3 H- 
rRNA we can say that about 50% of the 
gene region present in the rDNA before 
digestion has been recovered in the gel 
(Fig. 19a). Since there undoubtedly 
were losses during digestion, labeling 
with 32 P, and subsequent electrophoresis, 
this result indicates that a large fraction 
of the gene region is associated with 

Upon stronger digestion (Fig. 19 b and 
c) the gene region disappears rapidly, 
while a fraction of the 32 P-labeled DNA 
disappears more slowly. Since the DNA 
in this nucleolar preparation was 72% 
rDNA, it is probable that this more 
slowly degraded fraction is largely spacer 

The apparent presence of histones on 
actively transcribed ribosomal genes is 
somewhat surprising; especially so in 
view of the well-entrenched theory that 
histones function as repressors of genetic 
information. It is a sufficiently important 
observation that we intend to attempt 
to verify it as soon as possible by inde- 
pendent methods. If our present interpre- 
tation is correct, interesting questions 
are immediately raised concerning the 
manner in which RNA polymerase is 
able to traverse histone-covered DNA. 
The intriguing possibility is also raised 
that histones on active genes are different 
in conformation, secondary modification, 
or some other measurable parameter 
from histones on inactive DNA. This 
possibility is also open to experimental 

Iodination of Xenopus HlSTONE F 2a i 
in Chromatin 

S. Biroc 

Xenopus histone F 2a i has four tyrosine 
residues, all of which can react with 
radioactive iodine under appropriate 
conditions. However, when histone F 2a i 
is complexed in native chromatin, all 
four tyrosines are not equally reactive, 
suggesting that their individual reactiv- 
ity depends upon their accessibility from 

the external milieu and further suggest- 
ing that we can use reactivity to iodine 
as a probe of the three-dimensional ar- 
rangement of F 2a i in chromatin. 

It was first necessary to determine the 
location of the Xenopus F 2al tyrosines 
along the amino acid chain. Xenopus 
F 2al has not been sequenced; however, 
it appears to have a sequence nearly 
identical to the known sequence of calf 
thymus F 2iU . The molecules of both spe- 
cies co-migrate on acid-urea or SDS 
acrylamide gels. They have the same 
number of tyrosine residues. After iodi- 
nation and tryptic digestion they both 
yield identical iodinated tryptic peptide 
maps. Since, in general, F 2al is the most 
evolutionarily conserved protein that is 
known, we believe it is safe to use the 
amino acid sequence of calf thymus F 2al 
to order the tyrosines of Xenopus F 2al . 
Figure 20 shows the electrophoretic sepa- 
ration of the iodinated (tyrosine-con- 
taining) tryptic peptides from Xenopus 
F 2al and shows where they are located 
within the complete molecule. 

The influence of chromatin structure 
on the reactivity of each of these tyrosine 
residues was examined in the following 
manner. Chromatin from Xenopus cul- 
tured cells was iodinated in various con- 
centrations of NaCl ranging from zero 
to 2 M. Each mixture also contained a 
known amount of the free amino acid 
L-tyrosine to serve as an internal stand- 
ard. After iodination the specific activity 
of the L-tyrosine was measured as well 
as the specific activity of each of the 
four F 2al peptide tyrosines. The specific 
activity of each peptide tyrosine was 
divided by the specific activity of the 
free tyrosine to calculate the reactivity 
index of the peptide tyrosine. A reactiv- 
ity index of less than 1.0 indicated that 
the peptide tyrosine was less accessible 
to iodination than the free tyrosine, and 
a reactivity index greater than 1.0 indi- 
cated it was more accessible than free 
tyrosine. The reactivity indices of each 
F 2at tyrosine are plotted against NaCl 
concentration in Fig. 21. In very low 





¥1 >* 

- a 

T3 03 

-JS bD 

03 O 

.3 ^3 

T3 03 

.2 S 

re c3 

* fc 

5. 8 

© £ 

d g 

43 -d 

5 ° 

fa & 

P, += 

© 03 



d cu 

02 ■** 

l -a 

+= 03 

• d a. 

02 **• 


■S a 

T3 O 

.2 -£3 




o3 03 
03 CD 

w I 

02 CD 

2 •** 

O +3 

03 >. 

> a 


03 02 

CD -t-3 

S a 
o <D 

43 a 

ft °o 

■p p, 

o o 

-3 43 
9 *" 
"« -e 

.a 5 

0) bfi 
5-1 d 
3 'd 




9 ® 



05 CD 
'sU *-" 

> £ 


cd bJD +-> 

y3 "g a 
d <! w 

CD . cj 

Q d 

bC O 
03 £ 

CD 03 g) 













£-2 g 

02 ^3 

<D <L) "^ 

•i-H -rH 03 





ionic strength, where chromatin presum- 
ably is in its normal nondenatured con- 
formation, all four tyrosines were less 
reactive than free tyrosine. Tyrosines 98 
and 72 barely reacted at all, while tyro- 
sines 51 and 88 were about one-third as 
reactive as the internal standard. Pre- 

sumably all four are more or less "buried" 
within the chromatin structure. As the 
salt concentration is raised, two interest- 
ing observations can be made. (1) At 
0.5 M NaCl tyrosines 98, 72, and 51 re- 
main unaffected. However, tyrosine 88 
has increased to almost maximal reactiv- 



0.8 r 

0.5 1.0 1.5 2.0 

NaCI concentration 



Fig. 21. Reactivity of histone F 2a i tyrosines iodinated in chromatin in increasing NaCI 
concentrations. Xenopus chromatin was dissolved in (1) 5 ml Tris, 1 ml EDTA, (2) 0.3 M 
NaCI, (3) 0.5 M NaCI, (4) 0.75 M NaCI, (5) 1.0 M NaCI, and (6) 2.0 M NaCI, and then 
iodinated in the presence of an internal standard. The histones were extracted and electro- 
phoresed on acid/urea gels. The radioactive F 2a i was eluted from the gel piece, trypsin-digested 
and electrophoresed at pH 3.5 to determine how much radioactivity was associated with each 
tyrosine. The specific activity of each peptide tyrosine was compared to the specific activity 
of free tyrosine for each respective NaCI concentration. Reactivity index is plotted as a 
function of NaCI concentration for each F 2a i tyrosine. 

ity in this salt. Control experiments 
showed that histone Ft was completely 
dissociated from the chromatin at 0.5 
M NaCI, while the other four histones 
remained bound to the DNA. This sug- 
gests the possibility that tyrosine 88 is 
involved in interaction with histone F lt 
Alternatively, removal of F x may allow 
a conformational change in the chroma- 
tin that exposes tyrosine 88. (2) Raising 
the NaCI concentration to 2.0 M causes 
an increase in the reactivity of the other 
tyrosines so that at 2.0 M NaCI all four 
have undergone a five- to tenfold in- 

crease over their reactivity in low ionic 

These experiments demonstrate that 
the iodination reaction is a sensitive 
probe for detecting changes in the asso- 
ciation state (or conformation) of his- 
tone F 2a i- The fact that we are rinding 
histones on actively transcribed genes 
raises the likelihood that those histones 
may be in a different conformation than 
histones on inactive genes. In future ex- 
periments we hope to use the iodination 
reaction as a tool to search for such 
conformational differences. 




Y. Suzuki and P. Giza 

We have been studying the regulation 
of fibroin gene action in the silkworm 
Bombyx mori. Our goal is to understand 
which cellular constituents are responsi- 
ble for the selective transcription of the 
genes in the posterior silk gland. Our 
approach is to isolate chromatin, RNA 
polymerase, and whatever regulatory 
factors may occur in the posterior gland, 
and to attempt to reproduce faithful 
transcription of fibroin genes in vitro. 

Toward this goal, several studies have 
been initiated and are reported here: 
(1) morphological observations of silk 
gland nuclei during development; (2) 
efforts to prepare probes for detecting 
fibroin mRNA sequences synthesized in 
vitro; (3) pulse-labeling experiments in 
vivo; and (4) the isolation of nuclei and 
the synthesis of fibroin mRNA in vitro. 

An unexpected and encouraging ob- 
servation was the fact that in RNA 
preparations labeled in vivo for only 30 
minutes the synthesis of fibroin mRNA 
comprised 10% to 15% of the total 
radioactive transcripts in the late fifth 
instar. This quantitation is possible be- 
cause of the unique nature of the fibroin 
mRNA sequences {Year Book 73, p. 24). 
Ours is the first quantitative measurement 
of a single messenger RNA in pulse- 
labeled material by a direct chemical 
method. Similar results have been ob- 
tained with isolated silk glands in organ 
culture, indicating that the posterior silk 
gland is a self-regulatory organ for 
fibroin gene transcription in the late 
fifth instar. Although the silk gland 
nuclei possess a convoluted and ramified 
structure, which makes their isolation 
difficult, to say the least, the unusual 
predominance of fibroin mRNA synthesis 
in the pulse-labeled material has com- 
pelled us to attempt the fourth study, the 
isolation of nuclei and the synthesis of 
fibroin mRNA in vitro. This approach 
would narrow the localization of the 

regulatory factors for fibroin gene tran- 

Morphological Observations of Silk 
Gland Nuclei during Development 

P. Giza and Y. Suzuki 

Before attempting to isolate silk gland 
nuclei for transcription in vitro we felt 
that a better understanding of nuclear 
morphology was essential. It is known 
that silk gland nuclei undergo dramatic 
changes in appearance as the larva ma- 
tures, suggesting that the isolation of 
intact nuclei could present difficulties. 
Accordingly, silk glands were studied 
from late embryos (about 1.5 day before 
hatching) through the late fifth instar. 
All preparations shown in Fig. 22 are 
fixed, Feulgen-stained, whole mounted 
silk glands at the larval stages indicated. 

The shape of the nuclei from late em- 
bryos (not shown here) to early third 
instar is not unusual compared to nuclei 
of other eukaryotic cells. However, the 
silk gland nucleus is very much elon- 
gated and occupies a large portion of the 
cellular volume. During the middle to 
late third instar the nuclei begin to 
ramify at both ends of the elongated 
nuclear structure. The process continues 
through the fourth and fifth instar s and 
culminates in the extremely complex 
structure shown in Fig. 22f. This nuclear 
ramification is associated with signifi- 
cant increases in cellular DNA, a process 
termed "polyploidization." 

Pondering this complex structure (Fig. 
22f) one must ask, how are 400,000 sets 
.of genomes in the posterior gland cell 
in the fifth instar organized structurally 
and functionally? We do not see a typi- 
cal "polytene" structure. It has been 
generally accepted that the nuclei present 
an amorphous "interphase-like" appear- 
ance, but convincing data are scant. 
Therefore, we have studied the nucleus 



Fig. 22. Morphological changes in silk gland nuclei during development. They are all 
posterior glands except one middle gland (le). Preparations were fixed, stained by the Feulgen- 
Schiff reaction, dehydrated, and mounted in Harleco Synthetic Resin. The bars are 100 /xm. 
(a) first instar (2 days after hatching), (b) second instar (3 days after first ecdysis), (c) third 
instar (4 days after second ecdysis), (d) fourth instar (4 days after third ecdysis), (e) fourth 
instar (2 days after third ecdysis), (f) fifth instar (7 days after fourth ecdysis). Figure 22e 
is displayed to show hexagonal cell boundaries clearly. 

by high power microscopy after Feulgen 
staining. To our surprise we have ob- 
served a large number of Feulgen-posi- 
tive granules, somewhat like individual 
"chromosomes," in the polyploid nucleus 
(Fig. 23). The size and shape of these 
granules are quite similar to those of 
individual chromosomes in spermato- 
cytes. In fact, the overall staining of the 
nucleus appears to be due to the indi- 
vidual staining of thousands of these 
granules. These structures have been 
seen throughout the nucleus and have 
now been observed in embryonic through 
third instar nuclei. Whether they persist 
throughout larval development or are 

present only at certain periods is not 
known. We also do not know if the gran- 
ules are present in the fourth and fifth 
instars, because these nuclei are thick 
and stain darkly with little internal de- 
tail. Very preliminary results have re- 
vealed that the number of granules is 
increasing during silk gland develop- 
ment; the approximate number is a 
product of the polyploid number and the 
haploid chromosome number. These ob- 
servations and interpretations are remi- 
niscent of Geitler's early studies of 
endomitosis some 40 years ago. Recently, 
C. Thomas (this Report) has shown that 
the fibroin genes and ribosomal DNA are 



Fig. 23. Feulgen positive granules in silk- 
gland nuclei. Middle silk-gland nuclei of early 
first instar. The bar is 10 /tin. 

distributed randomly in the nucleus in 
all stages and in the middle as well as the 
posterior silk gland. It is also known 
that thousands of nucleoli are distributed 
throughout the nucleus. We can now ask 
whether the fibroin genes can be detected 
on single granules at a frequency of 1 out 
of 28 granules, thereby providing defini- 
tive identification of the granules as 
individual chromosomes, because B. mori 
has 28 haploid chromosomes which are 
round and indistinguishable from each 
other. We know, too, that there is one 
fibroin gene per haploid complement. 
Thus, the following picture of chromo- 
somal organization in the silk gland 
nucleus emerges. Multiple endomitotic 
replications of individual chromosomes 
are followed by random segregation of 
the replicated chromosomes, each indi- 
vidual chromosome maintaining its inde- 

Preparation of Assay Probes for 

Detecting Fibroin mRNA 

Synthesized in vitro 

Y. Suzuki 

In order to detect nascent or frag- 
mented fibroin mRNA in mixtures of 
various RNA species, we need hybridiza- 
tion probes like fibroin genes or comple- 
mentary DNA synthesized by reverse 
transcriptase. Although the optimal 
probe will probably be supplied by 
fibroin gene purification and amplifica- 
tion of the gene linked with a plasmid 
in bacteria' other approaches are being 

The poly rC-cellulose column that was 
described in Year Book 73, p. 24, has 
been found to be unsatisfactory. 

E. coli DNA polymerase is known to 
synthesize DNA from RNA templates. 
Therefore, using purified fibroin mRNA 
(>90% pure) as template and oligo dT 
as a primer, complementary DNA was 
synthesized. The cDNA was hybridized 
with radioactive total cellular RNA from 
the posterior gland, and from the hybrid 
thus formed, RNA was recovered and 
digested with RNase T x , and the digest 
was fractionated with DEAE-Sephadex. 
The profile was characteristic of a 50% 
pure fibroin mRNA, suggesting either 
that impurities in the template mRNA 
preparation were transcribed preferen- 
tially or that transcripts of noncoding 
regions of the mRNA make up about 
half of the short sequence synthesized. 
The latter possibility was verified by the 
following experiment. Reverse transcrip- 
tion was carried out using oligo dT- 
cellulose as a primer. The cDNA-oligo 
dT-cellulose thus obtained was packed 
into a jacketed column and hybridized 
with the radioactive total cellular RNA 
from posterior gland. The hybrid was 
eluted with thermal elution, and the 
radioactive RNA was tested for its Ti 
fingerprint. The profile was character- 
istic of pure fibroin mRNA. The result 
indicates that, although the reverse tran- 
script is composed of complementary 



sequences of noncoding and coding re- 
gions to fibroin mRNA, a longer fibroin 
mRNA sequence than the cDNA has 
been trapped, and the cDNA column is 
useful for purifying fibroin mRNA. 

It is known that fibroin genes bind 
more actinomycin D than the bulk of 
B. mori DNA and can be partially puri- 
fied by actinomycin D/CsCl centrifuga- 
tion. This technique has been employed 
for the quantification of fibroin genes 
and the structural analysis of the genes 
using purified fibroin* mRNA as an assay 
probe. We have asked whether the total 
cellular RNA of the late fifth instar can 
be used for hybridization and whether 
the RNA recovered from the fibroin gene 
region is pure fibroin mRNA. The Ti 
fingerprint of the RNA was identified as 
pure fibroin mRNA, indicating that the 
partially purified genes suffice to remove 
pure mRNA from the total cellular 
RNA. However, when the total cellular 
RNA of earlier stages of the fifth instar 
was employed, the fingerprint of the 
hybridization RNA showed some im- 
purities, imposing limits on usefulness 
of the technique. 

Fibroin mRNA Synthesis in a 
Pulse-Labeled Silk Gland 

Y. Suzuki 

It has been generally believed that in 
eukaryotes a specific single mRNA would 
be difficult to detect by direct chemical 
methods like base composition or se- 
quence analysis. We wondered whether 
this generalization held for fibroin 
mRNA too; fibroin genes comprise only 
0.002% of the B. mori genome, and, in 
fact, in our earliest studies fibroin mRNA 
was not detectable in pulse-labeled 
RNA, but was detectable only after sev- 
eral hours of incubation. However, the 
methods employed several years ago were 
sucrose gradient centrifugation of RNA 
and base composition analysis of the 
fractionated RNA, the sensitivity of 
which was not satisfactory. 

The quantitative data listed in Tables 

3 and 4 indicate that the number of 
fibroin mRNA molecules/gene accumu- 
lated by 6 days of the fifth instar is 
about four times that of rRNA mole- 
cules/gene; and that by 12 hours of 
labeling, mRNA synthesis is highly pre- 
dominant, about 16 times greater than 
rRNA synthesis. These results suggest 
that the mRNA is more rapidly synthe- 
sized' (and less stable) than rRNA and 
could possibly be detected by chemical 
methods in pulse-labeled materials. 
Therefore, using Bio-Gel column frac- 
tionation and sequence analysis (Year 
Book 73, p. 20) , fibroin mRNA synthesis 
in a pulse-labeled material has been re- 
investigated. Posterior silk glands were 
labeled in vivo or in organ cultures in 
vitro with 3 H-uridine or 32 P0 4 3- for 30 
min, and the RNA was fractionated by 
Bio-Gel column (Fig. 24). The peak, 

TABLE 3. The Posterior Silk Gland RNA 
Accumulated by 6 Days of the Fifth Instar 

Number of Number of 

Amount Molecules/ Molecules/ 

/ug/Cell Cell* Genef 

rRNA 3.90 1.08 X 10 12 1.17 X 10 4 

mRNA 0.17 1.82 X 10 10 4.55 X 10* 

* Molecular weight of 28S plus 18S rRNA is 
about 2.2 X 10 6 , and that of fibroin mRNA is 
5.7 X 10 6 . 

f B. mori genome carries 230 rDNA genes 
and one fibroin gene per haploid set. The cell 
is about 4 X 10 5 -ploid. 

TABLE 4. Predominant Synthesis of Fibroin 

mRNA over rRNA during Different Time 

Intervals in Late Fifth Instar 

Number of RNA Molecules/Gene 

Number of rRNA Molecules/Gene 

rRNA Fibroin mRNA 

30 min 



12 hr 



6 days 





30 " 




30 60 

Fraction number 

Fig. 24. Fractionation of RNA synthesized in the posterior silk gland in 30 min on a Bio-Gel 
A-50m column. Two pairs of posterior glands were dissected at 6 days of the fifth instar, and 
incubated with 3 H-uridine for 30 min at 26° C. RNA was extracted and applied to a Bio-Gel 
column (2.6 X 70 cm). A 0.3 ml aliquot out of each 5.4 ml fraction was taken and counted 
for its acid-insoluble counts. Circles, 3 H cts/min; solid line, A 2 «3o monitored by a Gilson 

which appeared in the void volume of 
the column, had about 26% of the total 
radioactivity. This includes any recov- 
ered fibroin mRNA and any digested 
with RNase T 1 . The resulting oligo- 
nucleotides were fractionated on DEAE- 
Sephadex (Fig. 25). The oligonucleotide 
profile shown in Fig. 25 is characteristic 
of fibroin mRNA with some impurities. 
From the approximate purity and the 
fraction of radioactivity in the void vol- 
ume peak of Fig. 24, it was calculated 
that fibroin mRNA synthesized during 
a 30 min pulse-labeling period comprises 
about 13% of the total radioactive tran- 
scripts. From these results we conclude 
that mRNA is more rapidly synthesized 
and is less stable than rRNA (Table 4), 
and that the regulatory factors for fibroin 
gene transcription reside in the posterior 
gland per se and not in the hemolymph. 

^> 3 




1 )[ 


.f 2 







i * 

W » 





ra c 

t ion number 

Fig. 25. RNase T x fingerprint of fibroin 
mRNA fraction from pulse labeled material. 
RNA was labeled in vivo with 32 P04 3 ~ for 40 min, 
extracted, and fractionated by a Bio-Gel A-50m 
column. The void volume peak from the column 
like the one in Fig. 24 was recovered and 
digested with RNase Ti, and the digest was 
fractionated on a DEAE-Sephadex A25 column 
(0.9 X 12 cm). Circles, 32 P cts/min; Roman 
numerals indicate the position of oligonucleo- 
tides by carrier RNA. 



Fibroin mRNA Synthesis in Isolated 
Silk Gland Nuclei 

Y. Suzuki and P. Giza 

Our morphological observations of 
nuclei suggested that third instar nuclei 
might be appropriate for the isolation 
of intact nuclei. After a few trials, we 
abandoned the idea because too little 
material was obtained to process further. 

Gentle disruption of the silk gland 
cells from the fifth instar larvae has 
been unsuccessful. 

Finally, the posterior silk glands from 
fifth instar larvae were homogenized 
with a Dounce loose-fitting homogenizer 
in the presence of 0.2% NP40, and cyto- 
plasmic materials were eliminated by 
repeated centrifugations in the presence 
of NP40 and one centrifugation through 
2.2 M sucrose. Typical examples of the 
nuclei isolated by this method are shown 
in Fig. 26. Although the originally con- 
voluted and ramified nuclei (Fig. 22) 
have been fragmented, the fragmented 
nuclei maintain their integrity to a sur- 
prising extent, and even branched struc- 
tures are seen (Fig. 26) . Chemical analy- 
ses have revealed that DNA recovery is 
about 60%, and more than 95% of the 
cytoplasmic RNA is eliminated. In the 
isolated nuclei, the ratios of DNA, RNA, 
and protein are 1.0:1.5:4.6, indicating 
that both RNA and protein contents are 
higher than those in the nuclei of other 

RNA synthesis was carried out with 
these isolated nuclei at 25°C, and the 
incorporation was linear for 30 min, 
gradually slowing, but still increasing 
even after 90 to 120 min. The RNA was 
extracted and fractionated on a Bio-Gel 
A-50 m column. Figure 27 shows both 
the radioactivity profile and cold RNA 
absorbance, which is mostly due to con- 
taminating cytoplasmic RNA. Relatively 
high molecular weight RNAs were syn- 
thesized in vitro, and about 35% of the 
radioactivity resides in a region of 
greater than 28S rRNA. The cold RNA 
profile showed that there was little RNA 
degradation during incubation. The frac- 

l" ri " ' "-""I 




Fig. 26. Isolation of nuclear fragments from 
extremely ramified nuclei of posterior silk 
gland. The nuclear fragments were isolated from 
the posterior glands at 3 days of the fifth instar 
by the method described in the text, and 
stained by the Feulgen-Schiff reaction. Three 
typical examples of these are shown. The bar 
is 100 /Am. 

tion in the void volume of Fig. 27 was 
recovered and analyzed for its RNase 
Ti fingerprint. The profile was char- 
acteristic of very impure fibroin mRNA. 
The fraction was also hybridized to the 
act D/CsCl fractionated DNA, and the 
Ti fingerprint of the RNA from the 
fibroin gene region was found to be char- 
acteristic of fibroin mRNA with some 



Fig. 27. Fractionation of RNA synthesized 
in isolated posterior silk gland nuclei on a 
Bio-Gel A-50m column. Nuclear fragments iso- 
lated from two pairs of posterior silk glands at 
6 days of the fifth instar were incubated with 
3 H-UTP for 30 min at 25°C. RNA was ex- 
tracted, passed through a Sephadex G25 column, 
and applied to the Bio-Gel column (1.6 X 73 
cm) . A 25 fd aliquot out of each 2.7 ml fraction 
was taken and counted. Circles, 3 H cts/min; 
solid line, A 2 eo monitored by a Gilson spectro- 

20 40 

Fraction number 

impurities. These results demonstrate in pulse-labeling experiments in vivo, 

that some fibroin mRNA has been syn- Therefore, the results suggest that we are 

thesized in fragmented nuclei. However, losing regulatory factors in vitro. Our 

the amount is less than that observed next objective is to recover them. 


CELLS OF Bombyx mori 

C. Thomas 

Radioactive iodinated silk fibroin mes- 
senger RNA and ribosomal RNA have 
been used as probes to localize their 
genes in tissue sections of Bombyx mori 
by hybridization in situ. 

Gene localization is made easier in 
silk gland cells by the polyploidization 
that occurs during larval development: 
At the end of the fifth instar, each 
nucleus in the posterior or the middle 
gland cells has about 0.5 X 10 6 haploid 
equivalents of DNA. 

This study was undertaken to deter- 
mine whether the location of a gene is 
related to its state of activity. Whereas 
ribosomal genes are active in all silk 
gland cells throughout larval life, the 
fibroin genes are active only in posterior 
silk gland cells during larval feeding 

The specificity of iodinated fibroin 

mRNA as a probe for the gene was 
tested by hybridization with DNA that 
had been fractionated in an actinomycin- 
CsCl gradient. 

In Year Book 73, p. 16, preliminary 
experiments suggested a preferential 
localization of fibroin genes in the pos- 
terior silk gland cells close to the nuclear 
membrane; ribosomal genes were ran- 
domly distributed throughout cells of 
the posterior and middle silk glands. The 
intracellular distribution of the fibroin 
genes was reconstructed by serial sec- 
tions of posterior and middle silk gland 
cells from the second to the late fifth 
instar. Feeding and molting periods were 
also compared. Whether fibroin genes 
were functioning (Fig. 28) or not (Fig. 
29), they were distributed randomly 
throughout the nucleus and had no obvi- 
ous association with any visible nuclear 



Fig. 28. Localization of fibroin genes in five serial sections of the posterior silk gland during 
the third instar. A complete nucleus is included within these sections; the nuclear outline 
of the preceding section is delineated in each section. Hybridization reaction included 0.1 
Aig/ml 125 I-mRNA (3 X 10 7 cpm//*g), 50 /ug/ml carcass RNA and 100 Mg/ml E. coli sRNA. 



Fig. 29. Localization of fibroin genes in four serial sections of the middle silk gland during 
the third instar. A complete nucleus is included within these sections. Same conditions as 
for Fig. 28. 



Bombyx mori 

J . Morrow 

The goal of this project is the isola- Total Bombyx DNA of molecular 

tion of the silk fibroin gene and the DNA weight 60 X 10 6 was digested by an 

adjoining it in the silkworm genome, excess of a particular enzyme. The DNA 

The approach is to purify DNA contain- was then fractionated on a gradient, and 

ing-the fibroin gene from the rest of the the fibroin gene fragments were located 

DNA of B. mori. Subsequently the small by hybridization to radioactive fibroin 

amount of purified DNA will be repli- mRNA. The rate of sedimentation in a 

cated in Escherichia coli after covalent sucrose gradient, relative to a marker of 

attachment to a bacterial plasmid. known molecular weight, was used as a 

Since the DNA encoding the structure measure of the molecular weight of the 

of fibroin protein has a molecular weight gene fragments. The amount of neigh- 

of about 9 X 10 6 daltons, a larger frag- boring DNA still joined to the gene was 

ment containing the gene must be iso- estimated from its equilibrium buoyant 

lated. The size chosen was 25 X 10 6 density in an actinomycin D-CsCl gradi- 

daltons. Bombyx DNA was sheared to ent. 

this size by stirring at a constant speed. The results indicate that the average 

The DNA molecules containing the molecular weight of gene fragments was 
fibroin gene were then separated from 7-8 X 10 6 after Hind III digestion and 
most other DNA by actinomycin D-CsCl that they contained a small amount of 
equilibrium centrifugation {Year Book neighboring DNA. It is probable that the 
73, p. 15). Three successive gradient Hind III enzyme either does not cleave 
separations gave a 27-fold purification at any of the sites or else cleaves at one 
of the gene with 33% yield. Purification site within the gene and near the ends 
was assayed by hybridization of the of the gene. After EcoRI digestion, the 
DNA to radioactive fibroin messenger gene fragments had an average molecular 
RNA. The results indicated that little weight of roughly 15 X 10 6 and retained 
more could be achieved by further repe- a large amount of neighboring DNA. 
tition of the actinomycin D-CsCl band- Tentatively we conclude that this en- 
ing. Other gradients do not appear prom- zyme cleaves once within the gene but 
ising for isolation of this gene. At this not within a considerable distance out- 
molecular weight, the fragments contain- side the gene. Results of similar experi- 
ing the gene are not separated from the ments indicate that the Hind II enzyme 
main band of Bombyx DNA in Ag+- cleaves near the ends of the gene but not 
Cs 2 S0 4 gradients. Previous work showed within it. 

that the same should be true of Hg 2 + - Isolation of the Hind III gene frag- 

Cs 2 S0 4 gradients. ments would not provide the desired 

The Bombyx DNA and plasmid DNA neighboring DNA sequences. The EcoRI 

must have complementary termini in gene fragments are also unsuitable for 

order to be joined to form a recombinant isolation for two reasons. Their buoyant 

plasmid. Complementary termini are density in actinomycin D-CsCl does not 

formed by cleavage of DNA with some permit extensive purification from other 

restriction enzymes (sequence-specific Bombyx DNA. Their large molecular 

endonucleases found in bacteria). For weight suggests that it would be difficult 

this and other reasons, digestion of the or impossible to join them to plasmid 

fibroin gene and environs by several re- DNA through EcoRI cohesive ends be- 

striction enzymes was studied. cause of the low probability of cycliza- 



tion of large DNA molecules. This 
problem may be overcome by use of 
longer cohesive termini. 

After 27-fold purification of the fibroin 
gene by actinomycin D-CsCl gradients, 
one DNA molecule in every 400 contains 
the fibroin gene. Present work is directed 
toward joining this partly purified DNA 

to a bacterial plasmid through synthetic 
poly dA and poly dT termini. These 
termini can be polymerized on the re- 
spective DNAs by the enzyme, terminal 
transferase. Isolation of the gene will 
then consist of finding the occasional 
recombinant clone containing the fibroin 


/. B. Dawid, C. Kaushagen, J. L. Ramirez, S. Ohi, and W. B. Upholt 

Our work has been centered on the 
physical mapping of the mitochondrial 
DNA molecule (mtDNA). In all meta- 
zoan animals this DNA is a circular 
molecule with a molecular weight be- 
tween 10 and 12 X 10 6 . Some of the 
functions of mtDNA have been eluci- 
dated : This DNA codes for one large and 
one small mitochondrial ribosomal RNA 
molecule (rRNA) which, together with 
a number of proteins, form the distinct 
mitochondrial ribosome. In addition, 
mtDNA codes for several smaller (4S) 
RNAs, most, if not all, of which are 
transfer RNAs (tRNAs) and are in- 
volved in mitochondrial protein synthe- 
sis. ' Together, rRNA and tRNA sites 
occupy about 20% to 25% of the mtDNA 
molecule. The remaining parts of the 
mtDNA may code for messenger RNAs 
for mitochondrial proteins. A site of par- 
ticular interest in mtDNA is the origin 
of replication. This site may be visual- 
ized in the electron microscope as a dis- 
placement loop (D-loop), in which a 
short third strand of DNA is duplexed 
with one of the parent strands, displacing 
the other strand (see Kasamatsu and 
Vinograd, Annual Review of Biochemis- 
try, 43, 695, 1974) . This D-loop structure 
provides a convenient reference point in 
mapping studies. 

Some technical approaches of particu- 
lar value to mapping experiments have 
been pursued (Year Book 73, p. 47), in 
particular the use of bacterial restriction 
endonucleases and the covalent attach- 
ment of ferritin to 4S RNA for the 

visualization of 4S RNA/DNA com- 
plexes in the electron microscope. 

Mapping Studies on Mitochondrial 

DNA of Xenopus laevis and 

Xenopus mulleri 

J. L. Ramirez 

Here the primary approach involves 
the use of restriction endonucleases to 
provide reference points for the mapping 
of mitochondrial genes and for the ori- 
gin of replication in the mitochondrial 
DNA. It is known that EcoRI makes 
two and three cuts in X. laevis and X. 
mulleri mtDNA, respectively. (Year 
Book 73, p. 51). The two X. laevis frag- 
ments constitute 87% and 13%, and the 
three X.' mulleri fragments are 43%, 
39%, and 17% of their respective 

Endonuclease I from Serratia mar- 
cescens (Sma I) makes a single cut in 
X. mulleri mtDNA and none in mtDNA 
of X. laevis, or in any other of seven 
animal mtDNAs tested. 

The position of the D-loop (which 
marks the replication origin) with re- 
spect to the restriction sites was deter- 
mined by electron microscopy. DNA 
samples were fixed with glyoxal, which 
blocks branch migration of the D-loop 
in the fragments ; then they were fully or 
partially digested with restriction endo- 
nucleases, and examined by electron 
microscopy. Maps obtained for the two 
mtDNAs are shown in Fig. 30 a and b. 
In X. Laevis (Fig. 30a) EcoRI makes 
one of its two cuts within the D-loop, 


7.2 *f 


Sma I 

Fig. 30. Physical maps of mtDNAs; (a) X. laevis, (b) X. mulleri. EcoRI sites are repre- 
sented by underlined numbers. The different regions are expressed as percent of the length 
of the total genomes. The distance from EcoRI site 3 to the Stria I site is 5.3% of genome 
length. The D-loop is schematically drawn at the top of each map, with the displacement 
strand shown as a thick line. 

producing "H-forms." These forms con- 
sist of the two EcoRI double-stranded 
fragments bound together by the single- 
strand part of the D-loop. 

In X. mulleri mtDNA the D-loop was 
mapped first in relation to the single 
Sma I site (Fig. 30b). Next, the D-loop 
was located in an EcoRI digest and 
found in the smallest fragment (17% 
fragment). Gel electrophoretic analysis 
showed that the Sma I site is within the 
largest EcoRI fragment (43% frag- 
ment) . From these observations the three 
EcoRI sites, the Sma I site, and the 
D-loop could be oriented with respect 
to each other (Fig. 30b). This map was 
confirmed by our analysis of D-loop 
containing mtDNA fragments derived 
by partial digestion with EcoRI. 

Physical Mapping of 4S RNA Genes 
on Mitochondrial DNA 

S. Ohi and I. B. Dawid 

We have continued to work toward 
the physical mapping of 4S RNA genes 

on mitochondrial DNA. Most, if not all, 
of the mitochondrial 4S RNA molecules 
are tRNAs which are specific for the 
mitochondrial protein-synthesizing sys- 
tem. Since the visualization of tRNA in 
a hybrid with DNA depends on the tag- 
ging of the tRNA with a bulky component 
like ferritin, we reported last year an 
effort to link covalently tRNA to ferri- 
tin through an RNA-Hg-S-ferritin bond 
{Year Book 73, p. 49). Although we ob- 
served effective coupling, the complex 
formed was not stable enough to permit 
further work; a model complex, tRNA- 
Hg-S-BI-lysine, degraded exponentially, 
the half-life being about 18 hr at 4°C. 
An alternative method to covalently 
link tRNA to ferritin has been reported 
by Wu and Davidson (Journal of Mo- 
lecular Biology, 78, 1, 1973), and we 
have successfully reproduced their 
method. The method makes use of the 
Schiff base reaction; tRNA 3'-dialde- 
hyde generated by NaI0 4 oxidation is 
reacted with epsilon-NH 2 groups of ly- 
sine in ferritin and the conjugate is sub- 



sequently stabilized by NaBH 4 reduc- 
tion. The purified complex was stable 
over several months under storage at 

Having the adduct in hand, we then 
moved to the mapping of 4S RNA genes 
on Xenopus laevis mtDNA. Purified H- 
strand prepared by the poly(IG) method 
was hybridized with the small and large 
rRNAs together with ferritin-coupled 
4S RNAs (fer-4S RNA). Excess ferritin 
and fer-4S RNA were removed by elec- 
trophoresis on a 3% agarose gel. In 
another experiment, the sample was 
simply diluted to reduce the background. 
Two heteroduplex molecules with inter- 
pretive tracings are shown in Fig. 31. 
The position of fer-4S RNAs is presently 
being mapped relative to the positions 
of the small and large rRNA duplex 
regions. It is already apparent that one 
of the 4S RNA genes is located between 
the small and large rRNA genes (see 
Fig. 31a). Similar observations have 
been reported by Wu, Davidson, Attardi, 
and Aloni in Hela cells (Journal of 
Molecular Biology, 71, 81-93, 1972). The 
mapping on the L-strand will be carried 
out in reference to the specific cutting 
site of the EcoRI restriction enzyme. 

A Comparative Study of the 

Mitochondrial Genomes of the 

Sheep and the Goat 

W. B. Upholt 

Several aspects of the mitochondrial 
genetic system can be studied by com- 
paring closely related mtDNAs. These 
include the mechanisms of the mainte- 
nance of the genetic information during 
evolution of a genome which is present 
in many identical copies, the correlation 
of the rate of nucleotide change of spe- 
cific sequences with the function of the 
sequences, and the role of nucleotide 
sequence and secondary structure in the 
control of expression and replication of 

Previous studies of mtDNA evolution 
in X. laevis and X. mulleri have shown 

that about 20% to 30% of the mito- 
chondrial genome, apparently that cod- 
ing for the mitochondrial ribosomal and 
transfer RNAs, is the most highly con- 
served portion. The remainder of the 
mitochondrial genome has diverged con- 
siderably more rapidly {Year Book 70, 
p. 44). To compare these more rapidly 
diverging regions in some detail, more 
closely related species are needed which 
have homologies in this region but which 
show measurable differences as well. 
After a comparison of published data 
for the relatedness of the nuclear genome 
of different species and preliminary com- 
parisons of different mtDNAs (see Year 
Book 71, pp. 22-24), the goat and the 
sheep were selected for further studies. 
Preliminary RNA-DNA hybridization 
experiments suggested that sheep and 
goat mtDNAs are more closely related 
than those of X. laevis and X. mulleri or 
of mouse and rat, but less related than 
the mtDNAs of subspecies of the mouse 
{Year Book 71, pp. 22-24). 

Goat and sheep mtDNAs were com- 
pared by an analysis of the fragments 
obtained after digestion with several re- 
striction endonucleases. EcoRI, which 
recognizes a sequence of six bases, gives 
five fragments with goat mtDNA and 
three with the sheep mtDNA. Two of the 
sheep fragments co-migrate with two of 
the goat fragments in agarose gels, indi- 
cating that three EcoRI sites in sheep 
and goat mtDNAs occur at identical 
sites. This was confirmed by mapping 
the order of the fragments as described 
below. When these two mtDNAs were 
digested with Hind III restriction endo- 
nuclease, which recognizes a different 
six-nucleotide sequence, two sites out of 
four were in common. Hae III endo- 
nuclease, which recognizes a tetranucleo- 
tide sequence, yielded 41 fragments larger 
than 200 nucleotide pairs; 13 of these 
appeared identical in sheep and goat 
mtDNA. The frequency of occurrence of 
common fragments in digests of these 
mtDNAs by these restriction enzymes is 
consistent with approximately 10% se- 


quence divergence between sheep and heteroduplex molecules show a single 

goat mtDNAs. heterology loop which is 1.7% of the 

The order of the EcoRI fragment in mitochondrial genome in length. The 

goat and sheep mtDNA (Fig. 32) has location of this loop was mapped rela- 

been mapped by electron microscopy tive to the EcoRI sites, and in this way 

according to the following procedure, relative to the D-loop. As shown in Fig. 

Goat mtDNA was digested with EcoRI, 33, the heterology region overlaps one 

and the mixture of five fragments was end of the D-loop. When the spreading 

denatured and annealed with denatured, was carried out in the presence of 60% 

intact (singly nicked) sheep mtDNA. formamide, the original heterology loop 

Heteroduplex molecules were formed be- increased in length, and a second loop 

tween one circular strand of sheep formed which appeared to overlap the 

mtDNA and between one to five goat other end of the D-loop. These observa- 

fragments. Double-stranded regions could tions show that the most rapidly diverg- 

be distinguished from single-stranded re- ing regions in sheep and goat mtDNA 

gions of DNA in the electron micro- overlap, or at least are very close to, 

graphs. The order of the goat fragments the two ends of the D-loop. 

could be determined from heteroduplex The sequence divergence between sheep 

molecules in which two to four of the and goat mtDNAs is at a lowlbut easily 

goat fragments were present. Since every measurable level, indicating that further 

goat EcoRI fragment has a distinct size, comparative studies of these DNAs 

each duplex region could be assigned to should provide interesting information on 

one or to a combination of several of the functional organization and evolu- 

these fragments. The single small region tion of mammalian mtDNA. 
of lowest homology between sheep and 

goat mtDNA (see below) was helpful as Mitochondrial DNA and RNA 

an additional reference point. This ap- in Drosophila 

proach, with modifications, can be used q Kaushagen 

to map the order of restriction fragments mi , , . ,. e ., , , . . 

in their parent molecule. . . The ch /^nz ft tioii of mitochondrial 

Replication of most mtDNAs begins "bosomal RNA (mt-rRNA) from Dro- 

o+ „ ™™fi„ „;+„ au n i \ u 4.u sophiia melanogaster (Year Book 72, p. 

at a specific site (the D-loop) by the A J[ , , y ,. v , ^ ' r 

unidirectional extension of a short dis- 43) . + has be , en . contlnue d- Ba f corn- 
placement fragment. In the sheep P^iianalysis was performed on «P 
mtDNA, the D-loop was found in ap- mt-rRNA prepared from Schneider's 
proximately 30% of the closed, circular C " ltured Drosophila melanogaster em- 
mtDNA molecules; it was 4.4% of the bry 7 c Cel } ] lne + n ° 32 2 > S ro ™ ln 3 ™J™ 
mitochondrial genome in length. The supplemented with «PO«. The "P RNA 
position of the D-loop was measured with Wi \ S P Unfied on a ;SUcro S e-SDS ^ 

respect to EcoRI site 1 (Fig. 32). ™£ then rUn ? n *£° ^ &* "f 

o i t i.i ! y°/° lormamide. These gels were sliced, 

Sequence homology between sheep and counted and th(j 32p RNA was eluted 

goat mtDNA was studied by heterodu- Aliquotg of thig RNA were h drol d 

plex mapping m the electron microscope. overnight in 0>3 A r Na0H at 3rc The 

Regions of low homology are visualized four nuc i eo tide S were separated by high 

as single-stranded loops, and homologous voltage paper electrophoresis, and the 

regions form duplex DNA. The degree percentage of radioactivity in each nu- 

of homology required for the formation cleotide was determined. The remaining 

of a stable duplex depends on the con- 3 2p mt -RNA eluted from formamide 

centration of formamide and salt in the gels was run on aqueous 2.4% acrylamide 

spreading solution. When spread from gels, eluted, and also analyzed for base 

50% formamide, sheep-goat mtDNA composition. Since these values were 




\ & 

# *'■«.#■ 

Fig. 31. Electron microgram and interpretive tracing of heteroduplexes formed by hybridiza- 
tion between mitochondrial fer-4S RNA, small and large rRNAs, and the H-strand of X. 
laevis mtDNA. Hybridization was carried out at 37 °C in the presence of 40% formamide 
for 2 hr at [Na + ] = 0.5 M. The concentration of each component during the incubation was 
as follows : DNA = 2 /ig/ml, fer-4S RNA = 40 /ig/ml, small rRNA = 2 ^g/ml, large rRNA 




' * ■ * ■* •* » **.4* 3- ■" ♦ ' ' ft* »*» *- • t # - ™ « ■ »**•>«,' *«<i * », ■ <*• • $* ^'« ** *■ r *-*"'* * ,* * , i r -t % - ' ~~ 

= 4 ,/ig/ml. The duplex regions between mtDNA and the large (Lg) and small (Sm) rRNAs 
are indicated. Closed circles indicate ferritin-4S RNA hybrids with the DNA, and open circles 
are probable fer-4S RNA/DNA hybrids. Ferritin background was reduced (A) by dilution, 
and (B) by gel-electrophoresis. Bar equals 0.5 X 10 8 dalton double-stranded molecular weight. 



Fig. 32. Composite map of EcoRI sites of 
the goat and sheep mtDNAs. Underlined num- 
bers (1-5) show the positions of the EcoRI 
sites in goat mtDNA. The sheep mtDNA is cut 
by EcoRI at sites 1, 4 and 5. Numbers inside 
the map show the percent of the entire goat 
genome in each of the goat fragments. At the 
top of the map, the sheep D-loop is represented 
and its position with respect to the EcoRI cut 
1 is shown. 

-i h 


-/ /- 

j : i i_ 

J I I L 

5 10 




Fig. 33. Positions of D-loop and region of 
lowest homology (heterology loop) between 
the sheep and goat mtDNAs in reference to 
EcoRI sites 1 and 2 in goat mtDNA. (A) 
diagrammatic representation of the position 
and size of the heterology loop; (B) diagram- 
matic representation of the position of the 
sheep D-loop. 

nearly identical to those determined for 
the RNA eluted from the formamide 
gels, the 32 P mt-rRNAs were considered 
homogeneous, and the base composition 
data were averaged (Table 5). 

The composition of these rRNAs is 
quite unusual in that it is lower in its 
content of guanylic and cytidylic acids 
(GC) than any other rRNA from mito- 
chondria or any other source studied 
previously. This result is interesting in 
two ways. First, the low GC content cor- 
relates with the unusual behavior of the 
Drosophila mt-rRNAs during gel electro- 
phoresis (see Year Book 72, p. 43). Sec- 
ond, there is a correlation between GC 
content of cytoplasmic and mitochondrial 
rRNAs in Drosophila. Drosophila cyto- 
plasmic rRNAs have a GC content of 
about 40% which is considerably lower 
than that of most other metazoan ani- 
mals. The mt-rRNAs in other animals 
have a GC content of about 40%, i.e., 
lower than the cytoplasmic rRNAs in 
the same animals and similar to Droso- 
phila cytoplasmic rRNA. In Drosophila, 
the mt-rRNAs have an even lower GC 
content, and the difference in composi- 
tion between cytoplasmic and mitochon- 
drial rRNAs is maintained. It will be 
interesting to look for a functional or 
evolutionary basis for these relations. 

In preparation for mapping these 
mt-rRNAs on the mtDNA of D. me- 
lanogaster, closed-circular mtDNA was 
prepared from embryos by ethidium 
bromide-CsCl gradients. Open-circular 
mtDNA was prepared from this material 
by singly nicking it with DNase. Dena- 
turation maps of nicked mtDNA were 
generated in an effort to establish one 
reference point on the circular DNA 
molecule. Denaturation was accomplished 
by incubation at high pli in the presence 
of 12.5% formaldehyde and 25 ml 
EDTA. The solution was then neutralized 
and spread for electron microscopy. De- 
naturations performed at several pH's 
revealed that one region representing 
about 25% of the molecule denatures 
completely even at the lowest pB. where 


TABLE 5. Nucleotide Composition of Large and Small mt-rRNAs of Drosophila 
(mole percent of nucleotides, mean ± standard deviation) 








% G + C 




7.0 ± 0.2 

8.4 ± 0.7 

17.5 ± 0.3 

18.4 ± 0.7 

40.5 ±1.2 
40.0 ± 0.9 
29.9 ± 0.9 

28.6 ± 0.2 

10.8 ± 0.1 
12.7 ± 0.4 
22.1 ± 1.1 
23.4 ± 0.8 

41.8 ±1.0 
39.1 ± 1.2 
30.5 ± 0.7 
29.5 ± 0.6 

17.7 ± 0.2 
21.0 ±0.4 
37.5 ± 1.2 
42.0 ± 0.6 


Lg and Sm refer to the large and small mitochondrial rRNAs (see Year Book 72, p. 43), 
and 28S and 18S are the cytoplasmic rRNAs. The base composition determined for the cyto- 
plasmic rRNAs are similar to those reported previously from several other laboratories. 

TABLE 6. Extent of Denaturation of D. melanogaster mtDNA 
at Various pH's (Mean ± standard deviation) 

Percent of 

Percent of 


Number of 

Length in 

in Largest 










32.1 ± 2.5 

26.4 ±1.7 



44.5 ± 7.1 

25.4 ± 2.5 



57.3 ± 5.7 

28.2 ± 4.1 



any denaturation occurs. This region re- 
mains the same length even at pH's 
where up to 60% of the molecule is de- 
natured in other regions (Table 6). This 
one large denaturation "bubble" will be 
used as a reference point for mapping 
the mt-rRNAs. 

Drosophila mtDNA has also been cut 

with Hind III and EcoRI restriction 
endonucleases, yielding three and four 
fragments, respectively. The molecular 
weights of the Hind III fragments were 
determined in the electron microscope; 
they are 5.24, 3.60, and 3.01 X 10 6 . An 
analysis of partially denatured restric- 
tion fragments is currently under way. 



R. E. Pagano, L. Huang, M. Takeichi, and K. Ozato 
with the technical assistance of E. Asch, D. Somerville, and W. Duncan 

The investigations being carried out 
in this laboratory are concerned with the 
dynamic movements cellular phospho- 
lipids and cholesterol can undergo in the 
membranes of mammalian cells and the 
effects perturbations in these motions can 
have on cellular activity. The basic 
premise of these studies is that these 

membrane components are not merely 
static structural elements of cells, but 
rather because of their ability to undergo 
dynamic movements, both within single 
cells and between contiguous cells, they 
are candidates for control mechanisms 
which participate in the regulation of 
cell growth and differentiation. 



To test the relation between membrane 
lipid dynamics and biological functions 
of cells, one needs, first, information 
about the spatial distribution of cell sur- 
face lipid components (and cholesterol). 
Taking advantage of some of the proper- 
ties of vesicle-cell interactions described 
below (see also Year Book 73, p. 52) and 
the technique of high resolution electron 
microscopic autoradiography, we have 
begun to determine the surface distribu- 
tion of one lipid component, phosphatidyl 
choline, on cultured Chinese hamster 
fibroblastic cells. 

If cell lipids and cholesterol levels can 
in some way regulate cellular activity, 
then a static picture of the patterns of 
distribution of various lipid components 
over cell surfaces is only a first step in 
understanding the mechanism. A second 
essential step is to learn if, and how, 
these membrane components can move 
between cells, and whether such move- 
ments result in significant compositional 
changes in the surface membrane of a 
given cell. Finally, it will be necessary 
to delineate the possible effects which 
alterations in the composition of cell sur- 
face lipids can have on regulating cellu- 
lar activity. For this purpose we have 
chosen to investigate the effects of cell 
surface lipid perturbations on lympho- 
cyte activation. 

Details of our studies on each of these 
topics are presented in separate sections 
which follow. 

Interactions of Phospholipid 
Vesicles with Mammalian Cells 

Studies of Mechanism 

R. E. Pagano and L. Huang 
with the assistance of E. Asch 

The use of artificially generated lipid 
vesicles (liposomes or phospholipid dis- 
persions) has recently been reported by 
several laboratories for producing a 
variety of modifications in the physiol- 
ogy of mammalian cells. These include 
the use of vesicles as carriers to cells of 
entrapped materials, promoting agents 

for cell fusion, and as tools for modify- 
ing the lipid composition of the mem- 
branes of intact cells. The possible 
molecular mechanisms by which lipid 
vesicles produce their varied effects, how- 
ever, have not been fully determined. 
Furthermore, many of the studies have 
been complicated by factors which make 
mechanistic interpretations difficult if not 
impossible. For example, some studies on 
drug entrapment have been carried out 
in whole animals or in serum-containing 
medium — conditions which could favor 
the reorganization of the lipid vesicles 
into other structures. Others have utilized 
lipid vesicles prepared by mechanical 
dispersion of a mixed lipid system in an 
aqueous phase — a technique which is 
known to produce a heterogeneous col- 
lection of multicompartmented structures 
of widely varying size and shape. 

In Year Book 73, p. 52, we presented 
our observations on the characteristics 
of the uptake of phospholipid by cul- 
tured Chinese hamster V79 cells upon 
incubation in a simple, balanced salt 
solution, with chemically and physically 
well-defined unilamellar lipid vesicles. In 
that report, detailed studies on the 
mechanism of this uptake were pre- 
sented, which demonstrated that the 
major pathway of lipid uptake by these 
cultured cells at 37 °C is a vesicle-cell 
fusion mechanism. In the present study, 
we show that an additional pathway, 
lipid exchange, may account for the re- 
maining lipid uptake. 

Chinese hamster V79 cells were grown 
in roller bottles in a culture medium 
containing radio-labeled palmitic acid. 
This method effectively labels the cellu- 
lar glycerolipids and phospholipids with 
tritium. After a sufficient growth period 
the monolayer of cultured cells was 
washed extensively with Gey's solution 
and incubated for one hour at 37°C with 
14 C-D0L (0.05 /xCi/mg) unilamellar 
vesicles (1 mg phospholipid/ml Gey's). 
Following this incubation, the super- 
natant was removed and briefly centri- 
fuged to remove any whole cells or frag- 



merits which were detached from the 
roller bottles during the course of the 
incubation. The supernatant from this 
centrifugation was then concentrated to 
a small final volume on an Amicon ultra- 
filtration device. The concentrated sus- 
pension was then applied to a Sepharose 
4B column and the effluent monitored by 
absorbance measurements at 280 nra, and 
determinations of 14 C-cpm and 3 H-cpm. 

In Fig. 34, the elution profile of the 
unilamellar vesicle fraction prior to in- 
teraction with cells is shown. It is seen 
to consist of a broad symmetrical peak, 
both in absorbance and in 14 C-cpm, cen- 
tered about fraction number 36. Fig. 35 
shows the profiles of absorbance at 280 
nm, 3 H-cpm and 14 C-cpm for a suspen- 
sion of 14 C-DOL vesicles following in- 
teraction with the 3 H-labeled cells. The 
profile of absorbance consists of four 
distinct peaks. The first peak, corre- 
sponding to the column void volume, 
represents a material of high molecular 
weight (± 20 X 10 6 daltons), e.g., cellu- 
lar fragments or lipoprotein aggregates, 

which were not removed from the vesicle 
suspension by low-speed centrifugation. 
The second peak, centered about frac- 
tion 36, represents the unilamellar lipid 
vesicles. Both peaks are seen to contain 
3 H-cpm and 14 C-cpm. The remaining 
two peaks of absorbance have no de- 
tectable amounts of radioactivity associ- 
ated with them and probably correspond 
to soluble proteins. In the control experi- 
ment in which the lecithin vesicles were 
not incubated directly with the cells but 
rather with a simple salt solution that 
had been in contact with the 3 H-labeled 
cells, a similar profile of absorbance vs. 
fraction number was obtained (Fig. 36). 
The first two major peaks are again seen 
to contain both 3 H and 14 C, but the rela- 
tive amounts are considerably different 
from those given in Fig. 35. Qualitatively 
similar results to those presented in 
Figures 34-36 were also obtained with 
cells whose lipids were radio-labeled by 
growth on 2- 3 H-glycerol. 

Fractions 35-48 from the column 
represented by the profile in Fig. 35 were 

35 45 





Fig. 34. Elution profile of unilamellar fraction of 14 C-DOL vesicles chromatographed on 
Sepharose 4B (2.5 X 40 cm). Absorbance at 280 nm, and "C-cpm vs. fraction number. 
















35 45 55 



4000 cl 




,i, l ,j. l i,.wr.jQ 

Fig. 35. Elution profile of unilamellar 14 C-D0L vesicles following a 1 hr incubation at 37 °C 
with Chinese hamster V79 cells. Cell lipids were labeled with 3 H by growth on 3 H-palmitic 
acid. Absorbance at 280 nm, 3 H-cpm, and 14 C-cpm are plotted vs. fraction number. 





0.1 - 






o I— — i I U 




35 45 55 




4000 °j 



Fig. 36. Cells were incubated for 1 hr at 37 °C in Gey's solution containing no vesicles. The 
supernatant was then used for a 1 hr incubation at 37 °C with 14 C-D0L vesicles and chromato- 
graphed on Sepharose 4B as in Fig. 35. 



pooled and the lipids subsequently ex- 
tracted with 2/l:chloroform/methanol. 
Analysis of the extracted lipids by thin 
layer chromatography (TLC) revealed 
six major spots (A-F, Figure 37). Spot C 
was identified as phosphatidyl choline, 
and F corresponded to cholesterol and 
other less polar lipids, e.g., triglycerides. 
The remaining spots were not firmly 
identified but are most probably lyso- 
phosphatidyl choline (spot A), sphingo- 
myelin (spot B), and glycolipids (spots 
D and E). A similar two-dimensional 
chromatogram was obtained from the 
lipids extracted from pooled fractions 
34-47 of the column represented in Fig. 

The spots from each thin layer plate 
were subsequently scraped from the TLC 
plate and analyzed for both 3 H and 14 C. 

The results of this analysis are given in 
Table 7. It is seen that all spots from 
both plates contain 3 H and 14 C, with the 
possible exception of the less polar lipids, 
spot F, which is very high in 3 H-cpm 
and contains practically no 14 C. Further- 
more, it is seen that the ratio 14 C-cpm/ 
3 H-cpm, of spots A through F varies 
considerably. If the ratio of the total 
14 C-cpm to total 3 H-cpm scraped from 
each plate is made, it is seen that this 
number (0.40 for TLC-I, and 1.22 for 
TLC-II) agrees well with the corre- 
sponding 14 C-cpm/ 3 H-cpm ratio of the 
samples which were applied to each plate 
(0.39 for TLC-I and 1.29 for TLC-II). 
These numbers also agree well with the 
ratios of the pooled samples prior to 
chloroform-methanol extraction, indicat- 
ing that no preferential extraction of 
the lipids occurred. 


f Vpl 

• — FFA 




^ c 

L "V 





x Or 


(2) C/A/M/HAc/H 2 

Fig. 37. Two dimensional thin-layer chroma- 
togram on silica gel of lipids extracted from 
14 C-labeled vesicles which were incubated with 
3 H-labeled cells (Fig. 35; fractions 35-45,). 
Chromatograms were developed in solvent sys- 
tems (1) chloroform-methanol-289r ammonia 
(65: 25: 5, v/v) and (2) chloroform-acetone- 
methanol-acetic acid-water (6: 8, 2: 2: 1, v/v). 
Six major spots were observed (A-F) ; the 
dashed lines show the position of known stan- 
dards (PC = phosphatidyl choline; PS = phos- 
phatidyl serine ; PE = phosphatidyl ethanol- 
amine; FFA = free fatty acids; LPL = less 
polar lipids, e.g., cholesterol, triglycerides) de- 
veloped under identical conditions. Or = origin. 

TABLE 7. Radioactive Analysis of Two- 
Dimensional Thin-Laver Chromatograms 
(TLC) of Lipids Extracted From 14 C-DOL 
Vesicles Following Their Incubation with 
3 H-labeled Chinese Hamster V79 Cells 


14 C-cpm/ 
3 H-cpm 14 C-cpm 3 H-cpm 

TLC-I: Vesicles Incubated Directly with Cells 





























Total from plate : 16052 




to plate: 

. . . 


Pooled samples 


extraction : 



Vesicles Incubated 

with Supematan 

(See text) 





























Total from plate: 6064 




to plate: 


Pooled samples 

before extraction : 




All of the observations cited above 
suggest that during the course of vesicle- 
cell incubation some of the cell lipids 
become an integral part of the phospho- 
lipid vesicles with no significant change 
in vesicle size. These results can best be 
explained in terms of an exchange process 
in which lipid molecules from the cell 
surface and vesicle are interchanged, 
with no net transfer of cellular lipid to 
the vesicles. The data in Fig. 36 show 
that when lipid vesicles are incubated 
with a balanced salt solution containing 
the lipids and proteins normally given 
off by cells placed in a protein-free 
medium, a smaller fraction of the cell 
lipids become associated with the lipid 
vesicles. Thus, the contribution of ex- 
change between lipids in the cell super- 
natant and the vesicles is small relative 
to the vesicle-cell lipid exchange. 

If an exchange is involved in the in- 
teraction of lipid vesicles with cultured 
cells, it is reasonable to assume that in 

vesicles comprised of several lipid com- 
ponents, these components might exhibit 
different rates of exchange with the cell 
membrane lipids. The use of mixed vesi- 
cles comprised of 14 C-cholesterol and 3 H- 
lecithin shows (see Fig. 38), however, 
that regardless of the mole fraction of 
cholesterol in lecithin, the two compo- 
nents enter the cell at the same rate when 
the vesicle-cell incubation is carried out 
at 37°C. Since only about 10% of the 
lipid uptake at 37°C can be accounted 
for by an exchange mechanism, while 
the remainder is most likely due to 
vesicle-cell fusion, this finding is not 
surprising. At 2°C, where fusion is largely 
suppressed, the differential exchange of 
lipids between vesicles and cells becomes 
important, with cholesterol entering the 
cell more slowly than lecithin. 

The actual mechanism of the vesicle- 
cell lipid exchange demonstrated here 
remains to be elucidated. The exchange 
process might be the result of one or 



— -I 1 1 1 




1 1 1 1 r 




10 20 30 40 50 60 10 20 30 40 50 60 



Fig. 38. Uptake of 3 H-dioleyl lecithin (circles) and 14 C-cholesterol (triangles) by V79 cells 
from mixed vesicles, containing either 22 or 46 mole % cholesterol. The ratio, 3 H-cpm/ 14 C-cpm 
in the applied vesicle suspension was 1.0. The ordinate is the observed cpm per 10 7 cells. 



more of the following: (1) a physical 
transfer of lipid components between 
membranes when cell and vesicle come 
into momentary contact; (2) an ex- 
change mechanism involving an enzy- 
matic transfer of acyl chains from the 
exogenously supplied lecithin to other 
plasma membrane lipids, and vice-versa] 
or (3) a carrier mechanism involving 
specific phospholipid exchange proteins 
which could transfer lipids between the 
vesicles and cells. These possibilities will 
be explored in future studies. 

A Proposed Mechanism for 
Vesicle-Cell Interactions 

The results of our previous studies on 
vesicle-cell fusion (Year Book 73, p. 52) 
as well as the present observations have 
led us to consider a two-step mechanism 
for vesicle-cell interactions as given in 
Fig. 39. This proposed mechanism can 
simultaneously account for both vesicle- 
cell fusion and some vesicle-cell lipid 
exchange. In this scheme, when a lipid 
bilayer vesicle collides with the plasma 
membrane of the cell (Fig. 39A), an 
intermediate structure (Fig. 39B), is 
formed in which the outer monolayer of 
vesicle lipids becomes an integral part 
of the outer monolayer of the plasma 
membrane lipid bilayer, with the aqueous 
interior of the vesicle and its surrounding 
inner monolayer of lipids splitting or 
dividing the bilayer of the plasma mem- 
brane. A similar type of structure has 
been postulated to explain the penetra- 
tion of echinoderm eggs by oil droplets 
studied in the 1930's by Chambers and 
Kopac. The driving force for generation 
of the intermediate structure shown in 
Fig. 39B would be the reduction in free 
energy of the system by elimination of 
the relatively high-energy lipid bilayer 
vesicle/water interface and its replace- 
ment with a lower-energy cytoplasm/ 
intermediate/water interface. During the 
lifetime of the intermediate, the outer 
monolayer vesicle lipids would be free to 
diffuse laterally in the plane of the 
plasma membrane and intermix with the 

Lipid Bilayer 
y Vesicle p 


7 in" 

Plasma Membrane 

1 1 . 1 1 1 1 

Fig. 39. Schematic diagram of proposed mech- 
anism for interaction of unilamellar vesicles 
with the plasma membrane of a cultured cell 
at 37 °C. The outer monolayer of vesicle lipids 
is represented by a heavy black line, and the 
inner monolayer, by a thin dashed line. (A) 
Lipid bilayer vesicle containing sequestered 
material collides with the cell membrane to 
form an intermediate structure as given in 
(B). In this structure, the outer monolayer of 
lipids from the vesicle may mix with the outer 
monolayer of lipids in the plasma membrane 
to give the intermediate structure shown in 
(C). Either intermediate (B or C) can revert 
to an intact vesicle, (A or D), or rupture as 
shown in (E) or (F). Rupture toward the cell 
interior would be recorded as a fusion event 
(E), whereas rupture toward cell exterior (F) 
would transfer all the vesicle lipids to the cell 
without any transfer of the trapped contents. 
Rupture is depicted only for the intermediate 
structure shown in (B). 

cell lipids as depicted in Fig. 39C. The 
intermediate structures in Fig. 39 B and 
C could pinch off to give intact vesicles 
containing the original complement of 
vesicle lipids (Fig. 39A) or vesicles con- 
taining cell lipids and vesicle lipids (Fig. 
39D). The latter case would be measured 
as a molecular exchange event. Accord- 
ing to this scheme, the cell lipids would 
be confined to the outer monolayer of 
the newly generated bilayer vesicle. The 



intermediate structures (Fig. 39 B or C) 
could also rupture. If the rupture occurs 
toward the cytoplasm (Fig. 39C), the 
discharge of the vesicular aqueous con- 
tents will be registered as a fusion event, 
and the vesicle lipids will be added sym- 
metrically to both sides of the plasma 
membrane. Such a process might be 
mediated by an enzymatic reaction, 
e.g., a phospholipase activity in the 
cytoplasm. If the intermediate structures 
rupture toward the outside, there will be 
no association of the trapped contents 
with the cell, and both the inner and the 
outer monolayer vesicle lipids will be 
added to the outer monolayer of the 
plasma membrane of the cell. 

The proposed mechanism is consistent 
with the following experimental observa- 
tions of vesicle-cell interactions. First, 
it was observed that while small mole- 
cules such as sucrose can leak out of the 
vesicle during its interaction with cells, 
no simultaneous leakout of metabolite 
(e.g., 3 H-deoxyglucose) from the cell 
occurs. This is contrary to observations 
of natural membrane fusions which have 
been demonstrated to be leaky processes. 
This discrepancy can be explained in 
terms of the proposed model by requiring 
that any transient leak formed during 
vesicle-cell fusion occur during the for- 
mation of the intermediate structure 
(Fig. 39B) and not during actual dis- 
charge of the vesicular contents into the 
cytoplasm of the cells. A second obser- 
vation, given in another report, demon- 
strating a lack of fusion of large (>1500 
A) multilamellar vesicles with cells, is 
also consistent with the proposed model. 
We argue that the intermediate structure 
(Fig. 39B) is unable to form in the large 
vesicles because of their relatively low 
surface energy. 

It should eventually be possible to 
further test the scheme proposed in Fig. 
39 if electron microscopic techniques of 
sufficient resolution directly reveal the 
existence of the postulated intermediate 
structures. The use of vesicles which have 
an asymmetric distribution of radio- 

labeled lipid may also prove useful. In 
the interim, it is hoped that the results 
presented in this study will prove bene- 
ficial, not only in dealing with the cellu- 
lar modification phenomena produced by 
lipid vesicles but also in considering 
natural membrane fusion processes. 

Interaction of Phospholipid Vesicles 
with Mouse Lymphocytes in vitro 

L. Huang, K. Ozato, and R. E. Pagano 

Last year we reported (Year Book 73, 
p. 83) that pretreatment of mouse lym- 
phocytes with phospholipid vesicle sus- 
pensions significantly enhanced the myo- 
genic activity of these cells in vitro. We 
have now examined this interaction in 
more detail. 

First, we investigated the kinetics of 
phospholipid uptake by mouse lympho- 
cytes. Vesicles comprised of 3 H-egg yolk 
lecithin (EYL) were prepared in Gey's 
salt solution by ultrasonic irradiation. 
The uptake of lipid by mouse (CBA or 
C57 Black) thymus or spleen cells was 
studied both at 37°C and 0°C. As shown 
in Fig. 40, microgram quantities of EYL 
became cell-associated in a 60 min in- 
cubation at 37 °C. The uptake was linear 
with incubation time and showed a 
strong temperature dependence. Spleen 
cells were more effective in uptake than 
thymus cells. 

In order to study the influence of lipid 
fluidity on the phospholipid uptake by 
lymphocytes, vesicles comprised of di- 
myristoyl lecithin (DML) were prepared 
and used for uptake studies. These vesi- 
cles have their phase transition tempera- 
ture (T c ) at about 25 °C as compared 
with those of EYL vesicles at — 10°C. 
Therefore, vesicles prepared from EYL 
are in a fluid liquid-crystalline state both 
at 37°C and 0°C. The result of an up- 
take experiment using DML vesicles is 
shown in Fig. 41. At 37°C, the amount 
of lecithin uptake by thymus cells was 
comparable to that by EYL vesicles 
(Fig. 40). In contrast to the findings 
with EYL vesicles, however, the uptake 






1 1 


— 1 -r ' 

Spleen, 37°C 






/ Thymus, 37°C 
/ • / 





y/ • 

a/ a 

Spleen, 0°C 

- /i 

__— — — - — ~ A — " 
o — 


1 _.. 

Thymus, 0°C 

! 1 ' L. .... 

10 20 30 40 50 60 

Incubation Time (mm.) 

Fig. 40. Uptake of 3 H-EYL (1.0 X 10 6 cpm/ 
mg) by mouse lymphocytes in vitro. EYL 
vesicle concentration = 1 mg/ml; 1 X 10 7 
cells/ml in Gey's solution. Circles: thymus 
cell; triangles: spleen cells. Open symbols: 
0°C incubation; filled symbols: 37°C incuba- 

of DML was greatly enhanced at 0°C. 
Therefore, we conclude that the lecithin 
uptake process is much more effective 
below the thermal phase transition of 
the exogenous lipid, a phenomenon also 
observed in uptake studies in which 
Chinese hamster fibroblastic cells (Year 
Book 73, p. 56) were used. 

In order to further characterize the 
uptake of phospholipid by lymphocytes, 
a series of experiments using EYL vesi- 
cles under different conditions was per- 
formed; the results are summarized in 
Table 8. Divalent cations such as Ca+ 2 , 
Mn+ 2 and Zn + 2 stimulated the uptake 
process at nontoxic concentrations 
(judged by the trypan blue dye exclusion 
test). Mg+ 2 ions, however, showed no 
effect up to a concentration of 5 mM. 
Low (mitogenic) doses of Concanavalin 
A (ConA) and succinyl-ConA reduced 

the incorporation of 3 H-EYL into thy- 
mocytes; lower doses, however, were less 
inhibitory. Inhibitors of energy metabo- 
lism (NaN 3 ) and monovalent cation 
transport (ouabain) only slightly re- 
duced the uptake, while an inhibitor of 
cellular motility (cytochalasin B) showed 
no significant effect. Fixation of cells by 
glutaraldehyde strongly inhibited the up- 
take process. Mild enzyme treatment of 
cells prior to incubation with vesicles 
exhibited various positive or negative 
effects. Proteolytic digestion with trypsin 
but not with pronase significantly re- 
duced the incorporation of phospholipid 
into thymocytes without cytotoxic effects. 
On the other hand, treatment of cells 
with phospholipases A and D resulted in 
enhancement of the uptake of EYL. 
Neurominidase treatment was ineffec- 
tive. In order to determine whether phos- 
pholipid vesicles are incorporated into 


T i 1 

1 1 ! 



37° C 

— • 


1 1 1 

i i 1 

10 20 30 40 50 
Incubation Time (min) 


Fig. 41. Uptake of 3 H-DML (4.5 X 10 6 
cpm/mg) by mouse thymocytes in vitro. 1 X 
10 7 cells/ml; 1 mg DML/ml in Gey's solution. 
Open circles: 0°C incubation, filled circles: 
37 °C incubation. 



TABLE 8. Effect of Mitogens, Metabolic 

Inhibitors, Divalent Cations, and Enzyme 

Treatment on the Uptake of 3 H-EYL 

by Mouse Thymocytes 

Uptake of 3 H-EYL 


(% of Control) 



ConA* (2 pg/ml) 


ConA (20 fig/ml) 


ConA (200 /ig/ml) 


Suc-ConA* (2 /*g/ml) 


Suc-ConA (20 /*g/ml) 


Suc-ConA (200 /*g/ml) 


NaN 3 t (10" 2 M) 


Cytochalasin B$ (10' 4 M) 


Ouabain^ (10" 3 M) 


Glutaraldehvde* (2%) 


Ca +2 § (5 mM) 


Mg +2 (5 ml) 


Mn +2 (0.5 ml) 


Zn +2 (0.1 ml) 


Trypsin (0.01%) 


Pronase (0.001%) 


Phospholipase A (0.01%) 


Phospholipase D (0.1%) 


Neuraminidase (0.0001%) 


* 1 X 10 7 cells/ml in Gey's solution were pre- 
treated with various reagents at 37 °C for 30 
min (15 min for all enzyme treatments). After 
two washes, cells were incubated with 3 H-EYL 
vesicle suspensions (1 mg/ml in Gey's solution) 
for 60 min at 37°C. 

t Present in pretreatment as well as during 
incubation with vesicles. 

% Present only during incubation with vesicles. 

§ All experiments with divalent cations were 
carried out in Ca +2 , Mg +2 -free Gey's (CMFG). 
Controls for these experiments were also in 

lymphocytes as intact structures, i.e., 
both phospholipid and the trapped 
aqueous content of the vesicles become 
cell-associated, we conducted the fol- 
lowing experiment. 

Vesicles containing 3 H-inulin trapped 
inside 14 C-EYL were prepared and used 
in an uptake experiment. As shown in 
Fig. 42, 3 H-inulin uptake matched nearly 
perfectly with that of 14 C-EYL at 37°C, 
strongly suggesting that intact vesicles 
are being incorporated into the cells. 
Since the incorporation was insensitive 
to an inhibitor of energy metabolism 

(NaN 3 ), it is unlikely that a process 
such as endocytosis is responsible for the 
vesicle uptake. Also, in electron micro- 
scopic autoradiograms of thin sections 
through vesicle-treated lymphocytes, 
neither attached vesicles nor radio-labeled 
lipids were localized exclusively on the 
cell surface. The possibility that intact 
vesicles adsorbed on the surface mem- 
brane of cells was therefore excluded. 
At present, fusion of phospholipid vesi- 
cles with lymphocyte plasma membrane 
seems most likely the major pathway for 
their incorporation at 37°C, although 
definitive . proof awaits further experi- 
mentation. At 0°C, uptake of trapped 
3 H-inulin was negligible (Fig. 42), ex- 
cluding any mechanism involving the 
incorporation of intact vesicles. 

Another possible mechanism for lipid 
uptake by vesicle-treated lymphocytes is 
the exchange of individual lipid mole- 
cules between vesicles and the cell sur- 
face. Such a possibility is supported by 
the following experimental results. When 
vesicles composed of 14 C-EYL and 3 H- 
cholesterol (1:1 molar ratio) were pre- 
pared and used in uptake experiments, 
the rate of entry into lymphocytes for 
3 H-cholesterol at 37°C was about twice 
the expected rate for cholesterol and 
EYL molecules had they been incorpor- 
ated into the cells together (Fig. 43). 
This finding suggests that in addition to 
vesicle-cell fusion, individual molecules, 
especially cholesterol, can be incorpor- 
ated into cells by an exchange mecha- 

Electron microscopic autoradiographic 
studies of mouse lymphocytes incubated 
with a variety of radioactively labeled 
lipid vesicles are in progress. An auto- 
radiogram obtained following 1 hr in- 
cubation at 37°C with 3 H-EYL vesicles 
is shown in Fig. 44A. The radio-labeled 
lipid is seen to be distributed into both 
the plasma membrane and internal struc- 
tures in the cell. This finding is in quali- 
tative agreement with biochemical analy- 
sis of fractionated cells. By contrast, 
autoradiograms of cells treated 1 hr at 







x 10 


° 8 

* 6 



I4 C-EYL rr^ 


d 4 




^ 2 





1 1 





































Incubation Time (min.) 

Fig. 42. Incubation of mouse thymocytes with phospholipid vesicles generated from "C-EYL 
(2.12 X 10 5 cpm/mg) and containing sequestered 3 H-inulin (500 mCi/g). Dashed curve repre- 
sents the expected amount of 3 H-inulin uptake calculated from the observed 14 C-lipid uptake 
and the 3 H-cpm/"C-cpm ratio of the applied vesicles (0.187). 2.5 X 10 7 cells/ml; 1 mg 
EYL/ml in Gey's solution. 




















ncuba t ion 




( min.) 

Fig. 43. Uptake of "C-EYL (2.4 X 10 5 cpm/mg) and 3 H-cholesterol (4.64 X 10 5 cpm/mg) 
by mouse lymphocytes from EYL vesicles containing 50 mole % cholesterol. Dashed curve 
represents the expected amount of 3 H-cholesterol uptake calculated from the observed "C-EYL 
uptake and the 3 H-cpm/"C-cpm ratio of the applied vesicles (0.75). 1 X 10 7 cells/ml; 0.5 mg 
EYL/ml in Gey's solution. 



\3>~) \i, >p»^i 

t Of 


3T3 *-.- 


C c 



Fig. 44. (A) EM autoradiogram of normal thymic lymphocyte incubated 1 hr at 37 °C 
with 3 H-EYL vesicles. (B) EM autoradiogram of normal thymic lymphocyte incubated 1 
hr at 37 °C with vesicles comprised of 1:1 EYL/ 3 H-cholesterol. Note accumulation of label 
on nuclear membrane. (C) EM autoradiogram of Chinese hamster V79 cell incubated with 
3 H-dimyristoyl lecithin vesicles 1 hr at 2°C. Thin sections were made parallel to culture dish 
on which cells were grown. (D) EM autoradiogram of Chinese hamster V79 cell incubated 
with 3 H-dimyristoyl lecithin vesicles 1 hr. Thin sections were made perpendicular to culture 
dish. Note accumulation of label at edge of cell (arrow). (E) Phospholipid transfer between 
contiguous cells. A polyoma virus-transformed baby hamster kidney cell (Py-BHK) was 
labeled with radioactive 3 H-sphingomyelin vesicles and allowed to grow on a tissue culture 
plate containing unlabeled mouse 3T3 cells. The grains over cells mark the sites of radio- 
active decay and approximate location of a labeled lipid molecule. (F) Same as (E) except 
the labeled and unlabeled cells are not in contact. Note lack of transfer between cells. 

37°C with vesicles comprised of 1:1 
EYL/ 3 H-cholesterol (Fig. 44B) show 
that the labeled cholesterol is largely in- 
ternalized, and resides almost exclusively 
on the nuclear membrane of treated cells. 
Cell fractionation studies to determine 
whether the exogenously supplied cho- 
lesterol has been appreciably metabolized 
to some other component are in progress. 

Spatial Distribution of Cell 
Surface Lipids 

M. Takeichi and R. E. Pagano 
with the assistance of W. Duncan 

The two-dimensional distribution of 
phospholipids and cholesterol in the plane 
of the cell membrane is thought to be 
governed by several general physical 



principles which have been elucidated in 
model systems and in studies of pro- 
caryotic organisms. These include the 
two-dimensional phase separation of 
membrane lipids, the presence of specific 
"boundary lipids" which surround spe- 
cialized membrane proteins requiring a 
particular environment, and packing 
constants which are imposed because of 
the existence of structures with very 
small radii of curvature. In order to test 
some of these ideas in animal cell mem- 
branes, which are known to be exceed- 
ingly complex compositionally as well 
as functionally, we are attempting to 
elucidate the spatial distribution of exo- 
genously supplied radioactive phospho- 
lipids introduced into cultured mam- 
malian cells by the techniques of vesicle 
interaction, described elsewhere in this 

Chinese hamster V79 cells were incu- 
bated with vesicles composed of 3 H- 
dimyristoyl lecithin for 1 hr at 2°C or 
37°C. The cells were then washed, fixed, 
sectioned, and prepared for electron 
microscope autoradiography. Figure 44C 
shows an autoradiogram obtained when 
cells labeled at 2°C were sectioned par- 
allel to the culture dish on which the 
cells were grown. The level is seen to be 
homogeneously distributed over the cul- 
tured cell surface. Similar results were 
obtained for cells labeled at 37°C, except 
that some internalization of the label 
occurred. In sections made perpendicular 
to the culture dish, however, an inhomo- 
geneous pattern of labeling was found in 
which the lipid label accumulated be- 
tween contacting cells (not shown) and 
at the edges of contact of individual cells 
with the substrate (Fig. 44D, arrow). 
Little of the lipid label was found on the 
upper surface of the cell exposed to the 
bathing medium or on the under, attached 
surface of the cultured cell. 

These observations demonstrate that 
the spatial distribution of one membrane 
lipid may not be random. Future experi- 
ments to determine the lateral distribu- 
tion of each of the major lipid classes, 

and the variation of this distribution 
with the degree of unsaturation of the 
fatty acid moiety of the phospholipid 
are in progress. 

Intercellular Exchange of Lipids 
M . Takeichi and R. E. Pagano 

The surface membranes of a wide 
variety of cell types have the capacity 
of forming junctions which are permeable 
to cytoplasmic constituents. Thus the in- 
tercellular exchange of soluble compo- 
nents is thought to be one type of inter- 
cellular communication which can exist 
in a collection of cells comprising a cell 
mass or tissue. We are interested in the 
possibility that membrane constituents, 
namely cellular phospholipids and cho- 
lesterol molecules, can exchange between 
cells when two cells come into contact. 
If lipid molecules were exchanged and 
mixed between two cells with different 
plasma membrane lipid compositions, the 
properties of the surface membranes of 
one or both of these cells could be modi- 
fied as a result of the exchange. This in 
turn might result in altered cellular 
activity if there is a relation between 
membrane dynamics and the biological 
functions of cells. The following experi- 
ments were done to obtain basic informa- 
tion about the existence of an intercellu- 
lar lipid exchange mechanism. 

Various types of cells (baby hamster 
kidney (BHK), polyoma virus-trans- 
formed BHK (PyBHK), mouse L cells, 
Balb/c 3T3 cells) were labeled by incu- 
bation with 3 H-sphingomyelin (SPHM) 
lipid vesicles. These cells were then 
seeded into nonlabeled cell cultures. 
Under the appropriate experimental con- 
ditions, the labeled cells adhered to the 
culture dish, and some of them made 
contact with nonlabeled cells. Such cul- 
tures were then fixed and examined auto- 
radiographically at the light microscope 
level. The autoradiograms of these cul- 
tures show that the 3 H-label is trans- 
ferred to originally nonlabeled cells 
within 60 min after mixing of the donor 



and recipient cells. Figure 44E demon- 
strates that lipid transfer has occurred 
between a labeled PyBHK and an un- 
labeled 3T3 cell which have come into 
contact. In Fig. 44F, where no cell-cell 
contact was made, no transfer of label 
took place. Experiments in which labeled 
and nonlabeled cells were mixed in a 
pellet demonstrated that no transfer of 
labeled 3 H-SPHM occurred, suggesting 
that junctions formed by normal bio- 
logical processes are required for the 
intercellular exchange of lipids, and that 
artificial physical contact of cells is in- 
effective for this purpose. 

Further experiments to clarify the 
biological significance of the intercellular 
exchange of membrane lipids and the 
mechanism by which this exchange is 
accomplished are in progress. Particular 
attention is being focused on the possi- 
bility that specific phospholipid exchange 
proteins at the cell surface may act as 
carriers for the movement of lipids 
between cells. 

Lipid Perturbations and 
Lymphocyte Activation 

K. Ozato and L. Huang 
with the assistance of D. Somerville 

In Year Book 73, p. 83, we reported 
that the pretreatment of T-lymphocytes 
with liposomes could enhance the mito- 
genic response to ConA. We have ex- 
tended our study to test the effect of six 
different lipid vesicle types on mitogenic 
responses in both T-cells and B-cells, 
in an attempt to delineate the relation- 
ship between the altered lipid composi- 
tion of the cells and mitogenic response. 
T-cell response to ConA was tested in 
cortisone-resistant thymocytes. B-cell 
response was tested in spleen cells using 
the B-cell specific mitogen LPS. Cells 
were exposed to liposome solutions for 
20 min at room temperature. Excess 
vesicles were removed, and cells were 
then stimulated with mitogens. Table 9 
shows the mitogenic response on day 2. 
We found that egg yolk lecithin (EYL), 

TABLE 9. Effect of Different Liposomes on 
Mitogenic Response 

3 H-TdR 





I. ConA* Control (RPMI 1640) 


EYL (1 mg/ml) 



(0.5 mg/ml) 



oleate (0.1 mg/ml) 


DPL (1 mg/ml) 


DML (0.5 mg/ml) 


DML (0.5 mg/ml) 


DLL (0.1 mg/ml) 


No ConA 


II. LPS Control 


EYL (1 mg/ml) 



(0.5 mg/ml) 



oleate (0.1 mg/ml) 


DPL (1 mg/ml) 


DLL (0.1 mg/ml) 




* Mitogenic responses to ConA or to LPS 
were tested using cortisone-resistant thymocytes 
or spleen cells, respectively. Cells were treated 
with different liposomes diluted in RPMI 1640 
at room temperature for 20 min. After washing, 
cells were resuspended in fresh medium. Then 
ConA (1 fig/ml) or LPS (5 /tg/ml) was added. 
3 H-TdR (1 fiCi/ml) was added to the culture 
2 days after plating and cells were labeled for 
18 hr. Each value represents mean of duplicate. 

EYL-cholesterol, EYL-cholesterol ole- 
ate, and dimyristoyl (DML) could en- 
hance the mitogenic response in both 
T- and B-cells when the optimal doses 
of pretreatment were performed. DPL 
(dipalmitoyl lecithin) and DLL (dilauryl 
lecithin) showed no appreciable enhance- 
ment. In the T-cell system DML showed 
the most significant enhancing effect 
(about threefold). In the B-cell system 
the cholesterol oleate showed remarkable 
enhancement, again about three to four 
times higher incorporation being noted. 
Liposomes did not change the optimal 
doses of mitogens when either suboptimal 
or superoptimal concentrations of mito- 
gens were employed. 



Biochemical analysis of the stimulated 
lymphocytes as well as electron micro- 
scopic autoradiographic studies are in 
progress to determine whether a correla- 

tion exists between the enhancement of 
the mitogenic response and the degree 
of internalization of the exogenously 
supplied lipids. 


D. M. Fambrough, P. N. Devreotes, A. K. Ritchie, and K. Tepperman 
with technical assistance of S. Cherry, A. Mabin, and W . Duncan 

Our interest for the past several years 
has been in the differentiation of the 
skeletal muscle plasma membrane and 
the organization of skeletal muscle fibers. 
In previous Year Books we have reported 
on the appearance and maturation of 
several plasma membrane properties re- 
lated to the specialized functions of 
skeletal muscle during development. We 
have also reported on some of the 
changes in plasma membrane properties 
which characterize denervated adult 
skeletal muscle and are reminiscent of 
the properties of embryonic muscle, and 
on some characteristics of abnormal 
muscle, namely in the muscular dysgenic 
mouse and in humans with the disease 
myasthenia gravis. Many of our studies 
have focused upon the acetylcholine re- 
ceptors of skeletal muscle, since they 
appear in the plasma membrane as 
readily detected functional molecules 
early in muscle development, are later 
confined to a small area of cell surface 
(the postsynaptic surface of the neuro- 
muscular junction), and in denervated 
skeletal muscle reappear all over the 
surface of each muscle fiber (as is the 
case in the embryonic state). Because 
acetylcholine receptors can be readily 
identified not only by their function but 
also through the use of a radioactive 
probe which binds to them with extreme 
tenacity (iodinated a-bungarotoxin) , 
they can be studied in ways which are 
unique. Thus we have been able to learn 
a great deal about their number and 
distribution in the plasma membrane in 
various physiological states, their rates 
of synthesis and destruction, and the 
mechanism by which they are incorpor- 

ated into the plasma membrane of mus- 
cle fibers and later removed and degraded 
by the fibers. Comparable data for other 
identified plasma membrane proteins in 
higher organisms are not yet available. 
In the past year we pursued our studies 
on the mechanisms of synthesis, incor- 
poration into plasma membrane, and 
degradation of acetylcholine receptors. 
The working hypothesis we had formu- 
lated in previous years includes the fol- 
lowing elements. Acetylcholine receptors 
are synthesized by the normal protein 
synthesizing machinery (although we do 
not yet know whether synthesis takes 
place upon membrane-bound or free 
polysomes) and their synthesis is sensi- 
tive to inhibition by puromycin and by 
cycloheximide. Newly synthesized re- 
ceptor components occur as organized 
macromolecular structures of molecular 
weight about 250,000 daltons (when 
solubilized in detergent solution) incor- 
porated into internal cell membranes (of 
unknown type and location). These "pre- 
cursors" are then inserted into the 
plasma membrane, the population of pre- 
cursors being enough to support two 
hours of new receptor incorporation into 
plasma membrane without any new pro- 
tein synthesis. Once in the plasma mem- 
brane the receptors are functional mole- 
cules and are free to interact with the 
radio-labeled probe, iodinated a-bungaro- 
toxin. Plasma membrane receptors are 
degraded by a random-hit process which 
is energy-dependent and proteolytic, 
probably involving internalization of 
membrane containing receptors and then 
degradation by lysosomal enzymes. Dur- 
ing the past year we developed techniques 



that will allow direct testing of several 
aspects of our model of the "life history 
of receptor molecules." We also devised a 
method for continuous monitoring of re- 
ceptor degradation in organ-cultured 
skeletal muscle and have begun extend- 
ing our findings to adult mouse muscle. 

We have also investigated another 
human disease, myotonic dystrophy, in 
which the muscle pathology suggested 
the possible participation of denervation. 
Since the appearance of extrajunctional 
acetylcholine receptors is a sensitive 
marker for denervated skeletal muscle 
surface, we have used our radio-labeled 
probe for acetylcholine receptors on 
biopsy specimens from patients with 
myotonic dystrophy. We have also ex- 
amined biopsy material from two pa- 
tients with the denervating disease 
amyotrophic lateral sclerosis. 

Eager to extend our studies to other 
specific membrane proteins, we have 
begun to probe for other skeletal muscle 
plasma membrane markers which might 
be used to investigate the metabolism of 
the plasma membrane. 

Characterization of Receptor Pools 

P. N. Devreotes and D. M. Fambrough 

Last year we described two classes of 
acetylcholine receptors in addition to the 
class functionally defined as surface re- 
ceptors. Our description was based on 
several observations. After a brief (20 
min at 0.1 ^g/ml) incubation of cultured 
chick embryo myotubes with a-bungaro- 
toxin, the cells are rendered physiologi- 
cally insensitive to ionophoretically ap- 
plied acetylcholine. After this treatment, 
however, there is a significant number of 
receptors which have not been complexed 
with a-bungarotoxin. These receptors are 
revealed by solubilization of the pre- 
treated myotubes in 1% Triton X-100 
followed by interaction of the detergent 
extract with 125 I-a-bungarotoxin. There 
are about 40% as many additional re- 
ceptors as there are rapidly saturable 
surface receptors. By several biochemi- 

cal criteria, which include the kinetics 
of interaction with a-bungarotoxin, pro- 
tection from a-bungarotoxin by 10 _3 ilf 
curare, and sucrose gradient sedimenta- 
tion, the additional receptors are indis- 
tinguishable from the rapidly saturating 
surface receptors. If one continues to 
incubate the intact myotubes with a- 
bungarotoxin, the percentage of addi- 
tional receptors revealed by Triton 
X-100 solubilization is eventually re- 
duced from 40% to 10%. Apparently, a 
class of receptors containing 30% as 
many receptors as does the surface in- 
teracts only slowly with extracellular 
a-bungarotoxin; while another class, 
containing 10% as many receptors as 
does the surface, cannot interact with 
a-bungarotoxin prior to extraction of the 
myotubes. The smaller class disappears 
rapidly in cells pretreated with 10 /xg/ml 
puromycin and is likely to be the pre- 
cursor of the surface receptors. The 
larger, slowly interacting class we have 
called "hidden" and its function is un- 
known. In cultured rat myotubes the 
two classes taken together are much 
larger than in chick and equal 100% as 
many receptors as there are rapidly 
saturable surface receptors. Our inter- 
pretation of the receptor classes and their 
possible relationships taken from these 
and other observations is summarized in 
Fig. 45. This model has been published 
in Journal of Cell Biology, 65, 335, 1975. 



Fig. 45. Hypothetical interrelationships be- 
tween receptor classes. 



Our goals in the past year have been 
to extend our understanding of the na- 
ture and relationships of the receptor 
classes denned above. Specifically, we 
have undertaken an electron microscope 
autoradiographic investigation of the 
"hidden" receptors. Perhaps the slow 
interaction with a-bungarotoxin is re- 
lated to their subcellular location. In 
addition, we have developed a technique 
to directly label the acetylcholine re- 
ceptor through its amino acid precursors. 
This -technique will soon be used to 
explore the interrelationships of the re- 
ceptor classes. 

In our initial discussion of the "hid- 
den" receptors it was stated that while 
the "surface" receptors are rapidly satu- 
rated with a-bungarotoxin the "hidden" 
receptors interact only slowly with extra- 
cellular a-bungarotoxin. Therefore, care- 
ful analysis of the kinetics of interaction 
of 125 I-a-bungarotoxin with intact myo- 
tubes should reveal two components. 
Figure 46 shows such an analysis. In this 
experiment a 30-fold excess of a-bungaro- 
toxin was used and therefore we would 
expect the kinetics to follow those of a 
pseudo-first order reaction. The data 
have been normalized so that 100 equals 
the number of surface receptors and 
therefore 130 equals the number of sur- 
face plus hidden receptors. Most of the 
receptors (about 100 on the normalized 
scale) interact with a-bungarotoxin rap- 
idly (tt/2 = 7 min), while the remainder 
(about 30 on the normalized scale) in- 
teract more slowly {t 1/2 = 1-2 hr). Also 
illustrated in Fig. 46 are the kinetics of 
interaction of 125 I-a-bungarotoxin with 
intact myotubes at 4°C. At 4°C only as 
many receptors as there are surface re- 
ceptors (100 on the normalized scale) 
interact with 125 I-a-bungarotoxin even 
after very long interaction times (24 hr) . 
This experiment suggests that at 4°C the 
hidden class of receptors does not bind 
a-bungarotoxin. Accordingly, we defined 
as another criterion for surface receptors 
that the receptors interact with a-bunga- 
rotoxin at 4°C. 

To determine whether the slowly in- 
teracting hidden class has a unique sub- 
cellular location, we prepared autoradio- 
grams on thin sections of myotubes and 
examined cells labeled at 4°C and at 
37°C. We predicted that at 4°C all the 
grains would represent receptors in the 
plasma membrane and thus be located 
at the cell perimeter. We predicted that 
at 37°C, in addition to the grains at the 
perimeter an extra 30% would have an- 
other subcellular localization. Unex- 
pectedly, in myotubes saturated with 
125 I-a-bungarotoxin at 37°C only a small 
fraction (~30%) of the acetylcholine 
receptors are associated with the perim- 
eter of the cell (Fig. 47). From this ob- 
servation it must be concluded that the 
rapidly saturating receptors we have 
functionally defined as "surface" are not 
located exclusively at the cell perimeter 
since, if they were, about 75% of the 
grains should be located there. For myo- 
tubes saturated at 37 °C we examined 
more than 1200 grains over 14 different 
cells in two independent preparations, 
and 55%-86% of the grains have not 
been located at the cell perimeter. Al- 
though we do not have quantitative data 
at this time, we have determined quali- 
tatively that in autoradiographs of cells 
labeled at 4°C many of the grains are 
also not located at the cell periphery. 
The unexpected distribution of "surface" 
receptors has complicated our analysis 
of the autoradiograms. Since many of 
the "surface" receptors can be located in 
what appears to be the interior of the 
cell, we would like to label the surface 
of the cell by an independent method. 
At present, several dyes and stains which 
would nonspecifically label the external 
surface of the plasma membrane are 
under consideration. 

Figure 48 shows the kinetics of dis- 
appearance, during puromycin treat- 
ment, of the class of receptors we have 
suggested to be the precursor of the 
surface receptors. Also shown is the 
simultaneous incorporation of new re- 








— 10 

i i 1 1 r 


12 3 4 5 


Fig. 46. Kinetics of a-bungarotoxin binding. Large sets of identical 5-day cultures were 
equilibrated in growth medium at 4°C and at 37° C. At each of the time-points ^I-a-bungaro- 
toxin was added to four cultures so that the concentration of a-bungarotoxin was 0.2 /Ag/ml. 
The time indicated is the time of preincubation of each set of four cultures with a-bungaro- 
toxin. At the termination of the experiment all the cultures were rinsed by immersion in a 
large bath of wash medium (Hanks 0.5% BSA, pH 7.2). The cultures were then extracted 
with 1% Triton X-100 and assayed for radioactivity in a Packard Auto-Gamma Scintillation 
Spectrom'eter. The value expressed is the mean of the four cultures in each set. The experi- 
ment was repeated three times. 

ceptors into the surface class. At present, 
these matched kinetics are the strongest 
evidence supporting the conclusion that 
these receptors are precursors of the 
surface receptors. Because of the indirect 
nature of this experiment, we wish to 
directly label the precursor and follow 
the incorporation of the label into the 
surface. In addition, an important ques- 
tion concerning the mechanism of in- 
corporation of new proteins into the 
surface membrane might be resolved by 
directly labeling the receptor. In Fig. 48 
the kinetics of disappearance of the 
precursor and the appearance of new re- 
ceptors appear to follow linear, zero- 
order kinetics. If this were accurate, it 

would imply an assembly-line type 
mechanism such that a newly synthesized 
receptor would spend a 2 hr transit time 
before its incorporation into the surface. 
On the other hand, if the kinetics were 
first-order exponentials, it would imply 
that a newly synthesized receptor would 
have equal probability of being incor- 
porated into the surface as a receptor 
which had been synthesized 2-3 hr ear- 
lier. We believe our puromycin kinetic 
data are not sufficiently accurate to 
resolve the linear and exponential models. 
However, if the receptor were directly 
labeled through its amino acid precursors 
and some labeled receptors appeared on 
the surface prior to 2 hr, the linear 






*-# : f 




f « 

Fig. 47. Electron microscope autoradiograph of receptor distribution in myotube. Several 
5-day myogenic cultures were labeled with ^I-a-bungarotoxin for 6 hr at 37 °C. After rinsing 
to remove unbound 125 I-a-bungarotoxin, the cultures were fixed overnight in 2% gluteralde- 
hyde at 4°C. The cultures were osmicated, stained in uranyl acetate, dehydrated, and 
embedded in Epon. Thin (gold interference color) sections were prepared and collected on 
Formvar carbon-coated grids. The grids were coated with a thin film of Ilford L-4 nuclear 
track emulsion and exposed for about a month at 4°C. Grids were developed in Microdol-X. 
Background was about 0.2 grain/100 n 2 . 

assembly-line model would be ruled out. 
To label the acetylcholine receptor 
with radioactive amino acids seemed im- 
practical from two points of view. Large 
quantities of high specific-activity amino 
acids would be expensive, and complete 
purification of the labeled receptor would 
be necessary before its radioactivity 
could be rigorously verified. We decided, 
therefore, to band the receptor by equi- 
librium density sedimentation and then 
to look for a shift in density of newly 
synthesized receptors following incor- 
poration of amino acids containing stable 
isotopes. Amino acid mixtures contain- 

ing the stable isotopes 2 H, 13 C, and 15 N 
can be obtained from Merck, Sharpe and 
Dohme. In addition, since the position 
of the receptor in the gradient is moni- 
tored by specified 125 I-«-bungarotoxin 
binding, no purification step is necessary. 
Metrizamide, a tri-iodinated benzamido 
derivative of glucose, provides a con- 
venient medium to density-band the 
receptor, since it does not disrupt the 
a-bungarotoxin-receptor interaction at 
concentrations with a density of up to 
1.5 g-m _H . Our preliminary attempts to 
shift the density of newly synthesized 
receptors with 2 H amino acids and to de- 



Puromycin Kinetics of Pool after overnight treatment with 

New surface 

Hours in Puromycin 

Fig. 48. Kinetics of pool disappearance and new receptor appearance. Decrease of "precursor" 
in presence of puromycin. All cultures were incubated overnight with 0.5-1.0 fig/ 'ml unlabeled 
a-bungarotoxin. The next day sets of 5-6 cultures were pretreated with 20 fig/m\ puromycin 
for the indicated times. All cultures were then washed to remove unbound a-bungarotoxin. 
In one experiment (open circles), sets were immediately extracted in 1% Triton X-100. In 
the other two experiments (filled circles, open triangles) the cells were scraped from the 
dish, homogenized by 25 strokes of a Dounce (pestle B) homogenizer (in 10 mM Tris 
at 0°C), and centrifuged at 50,000 g X 30'. The pellets were extracted with 1% Triton X-100, 
10 mM Tris, pH 7.8. Each of the extracts was incubated with 0.01 fig/m\ 125 I-a-bungarotoxin 
for 1 hr at room temperature. The complexes formed were freed of unbound a-bungarotoxin 
by Bio-Gel P-60 chromatography. The fractions in the excluded peak were pooled, and an 
aliquot was taken for analysis by velocity sedimentation in sucrose gradients. In all cases 
the area under the 10S peak in the gradient was measured. For each experiment, all values 
were normalized so that the control value equaled 0.10 of the surface. Incorporation was 
measured after cells were first saturated with unlabeled a-bungarotoxin for about 15 hr. 
The actual plateau level reached in this experiment was about 8% of the total surface, but 
it has been normalized to 0.10 for comparison with the pool decrease. 

tect the density shift in spontaneously 
formed metrizamide gradients were suc- 
cessful. The density shift produced, how- 
ever, was quite small (6-drop separation- 
s-drop half-width of bands). Shortly 
thereafter, another group reported equi- 
librium density sedimentation studies of 
bacterial enzymes in metrizamide gradi- 
ents. The authors observed that since 
metrizamide, being highly viscous, ap- 
proaches its equilibrium concentration 
gradient very slowly, it is possible to 
preform a stable metrizamide gradient 

and to centrifuge at very high fields 
(350,000 g) in order to rapidly equili- 
brate the macromolecular species (in 
our case, the acetylcholine receptor). In 
our initial attempts using a preformed 
metrizamide gradient (dp/dr = 0.03 
g-cm -4 ) we observed greater separations 
of "heavy" and "light" receptors (see 
Fig. 49a). By appropriate manipulation 
of the preformed gradient (dp/dr = 
0.015 g-cm -4 ) it should be possible to 
double this separation. In addition, we 
have shifted the density of receptors by 



incorporation of amino acids containing shifts seen in Fig. 49 a and b, it should 

13 C (see Fig. 49b). Since we can pur- be possible to observe large separations 

chase amino acids substituted with all of "heavy" and "light" receptors (40- 

three heavy isotopes ( 2 H, 13 C, 15 N) on drop separation-40-drop half-width of 

the same amino acid, and thus add the bands). 



t 07 









LU 0.3 




30 40 50 






40 50 


Fig. 49. Density shift of newly synthesized receptors. Several large 5-day cultures were 
incubated with 0.2 /Ag/ml unlabeled a-bungarotoxin overnight. After 16 hr of incubation the 
medium was switched to one containing amino acids substituted with 2 H (a) or 13 C (b) 
(preparation of the medium is described below) but still containing 0.2 fig/m\ unlabeled 
a-bungarotoxin. After 5 more hours of incubation the cultures were washed to remove 
unbound a-bungarotoxin and incubated in "heavy" medium for an additional 2 hr. During 
the final 30 min of incubation 125 I-a-bungarotoxin was added to the cultures such that the 
concentration was 0.2 /xg/ml. The cultures were washed at 4°C, the cells scraped from the 
culture dishes, homogenized in 10 mM Tris, and centrifuged at 17,000 rpm for 45 min. The 
pellet was extracted in a small volume of 1% Triton-10 mM Tris. An aliquot of the extract 
was mixed with a marker of 131 I-a-bungarotoxin-receptor complexes which were extracted 
from an independent set of cultures. The mixture was loaded in a thin band on a step 
metrizamide gradient containing 0.95 ml of each of five concentrations: 407c, 35%, 30%, 
25%, 20%) (W/W) of metrizamide. The solvent for the metrizamide was deuterium oxide. 
Centrifugation was carried out in a Beckman SW65 Rotor for 15 hr at 54,000 rpm at 4°C. 
Gradients were collected by puncturing the bottom of the tube. One hundred four-drop 
fractions were collected from each gradient into scintillation vials. Radioactivity was moni- 
tored simultaneously on appropriate channels for ^ 5 I and 131 I. After correction for crossover 
between the channels the data were plotted, and smooth curves were drawn through each 
peak (peak fractions contained at least 800 cpm). Data were then normalized to the highest 
point on the smooth curve; some points are therefore greater than one. Heavy medium was 
our normal growth medium except that 16 of the amino acids were substituted by an 2 H or 
13 C amino acid mixture and serum normally added to the medium was exhaustively dialyzed 
against Hanks. Amino acid mixtures were obtained from Merck, Sharp and Dohme and were 
either 98% atom purity 2 H or 80%o atom purity 13 C. 



Turnover of Receptors in Denervated 
Adult Muscle 

D. M. Fambrough and P. N. Devreotes 

Last year we reported that the rate 
of degradation of a-bungarotoxin bound 
to acetylcholine receptors in denervated 
rat diaphragm muscle in organ culture 
was a first-order exponential process 
with a half-time of about 24 hr. It was 
not possible to establish the identity 
between degradation of toxin and degra- 
dation of receptor, as was done for chick 
muscle in tissue culture, because of the 
much greater technical difficulties in 
working with adult muscle. Nevertheless, 
by analogy with cultured muscle, it 
seems almost certain that degradation 
of toxin is associated with degradation 
of receptor in adult denervated skeletal 
muscle as well. This hypothesis would 
be strengthened if it could be shown that 
the degradation of a-bungarotoxin bound 
to extrajunctional acetylcholine recep- 
tors in adult muscle were similar to 
degradation in tissue cultured muscle in 
its sensitivity to various metabolic in- 
hibitors and other environmental condi- 
tions. Exploration of factors influencing 
the degradation of a-bungarotoxin in 
denervated adult skeletal muscles was 
begun, but the rat diaphragm proved to 
be a poor experimental material because 
it cannot be isolated as a viable muscle 
without also including a large amount 
of other (intercostal) muscle and por- 
tions of several ribs, which serve as the 
area of insertion of the diaphragm onto 
the body wall. Since these tissues also 
bind a-bungarotoxin, they contribute to 
the observed degradation in a way that 
is difficult to control from one prepara- 
tion to the next and also difficult to 
evaluate for any given preparation. Thus 
a thin, easily denervated and easily dis- 
sected skeletal muscle with insertions 
onto long tendons was sought. The mouse 
extensor digitorum longus (EDL), a 
muscle in the lower hindlimb, was se- 
lected. This muscle has been used in 
many denervation and reinnervation 

Our experiments on acetylcholine re- 
ceptor metabolism in tissue culture uti- 
lized the constancy of total receptor 
number from one culture dish to the 
next, and often 20 to 50 culture dishes 
would be used in constructing a kinetic 
curve. When working with adult skeletal 
muscles it is not possible to dissect out 
20 to 50 muscles simultaneously; that 
would take several hours and require the 
sacrifice of many animals. Therefore, 
the next step in investigating receptor 
metabolism in adult skeletal muscles was 
to devise a method for obtaining an en- 
tire kinetic curve from one or a few 
muscles. We have not yet developed 
means to measure accumulation of acetyl- 
choline receptors or incorporation of new 
receptors into plasma membrane in single 
adult muscles. Such measurements are 
complicated by the long time needed for 
saturation of a muscle with a-bungaro- 
toxin and removal of the unbound toxin. 
However, we have been able to measure 
degradation of toxin previously bound 
to acetylcholine receptors continuously 
for periods of 24 hr or longer. This was 
accomplished by maintaining the muscle 
in a perfusion chamber which fits into 
the well of a gamma detector. The 
gamma radiation produced by decay of 
125 I, which is a measure of the number 
of a-bungarotoxin receptor complexes in 
the muscle, is continuously monitored. 
When iodinated a-bungarotoxin is de- 
graded, the resultant 125 I-tyrosine dif- 
fuses out of the muscle and is carried 
out of the gamma well by the perfusion 
solution. The loss of radiation from the 
muscle is due almost exclusively to 
degradation of a-bungarotoxin. Since 
degradation is very sensitive to tempera- 
ture, the gamma well and the perfusion 
solutions and associated pumps are 
housed in an incubator to maintain a 
constant 37°C. Some details of the per- 
fusion chamber are illustrated in Fig. 50. 
Figure 51 illustrates the loss of radi- 
ation from 8-day denervated mouse EDL 
muscle after exposure to 125 I-labeled a- 
bungarotoxin. The curve consists of three 





I — 2 /C0 2 95/5 

Fig. 50. Perfusion chamber for the growth of 
adult mouse muscle. The chamber consists of 
a glass test tube and silicone rubber stopper 
through which pass stainless steel tubes for 
delivery and removal of medium and oxygen. 
The muscle is pinned to a stainless steel grid 

first-order exponentials. The first repre- 
sents the washout of free bungarotoxin 
from the extracellular spacer of the mus- 
cle. The second has a total duration of 
about 5 or 6 hr and represents the loss 
of a variable but always rather small 
fraction of the total radioactivity. Its 
meaning is not known. Finally, the long 
third exponential represents the degra- 
dation of bound a-bungarotoxin and re- 
lease of the radio-labeled degradation 
product from the muscle. Some assump- 
tions are required for calculation of the 
half-time for degradation of a-bungaro- 
toxin. The major assumption is what 
fraction of the total radioactivity is 
associated with this exponential. It is 
known that the a-bungarotoxin associ- 
ated with acetylcholine receptors that 

occur at the former neuromuscular junc- 
tion is degraded at a much slower rate 
than is extrajunctionally bound a- 
bungarotoxin. In the mouse EDL the 
extrajunctional acetylcholine receptors 
constitute roughly 60% to 90% of total 
receptors during the first few weeks fol- 
lowing denervation. Another portion of 
the bound a-bungarotoxin is associated 
with material that cannot be solubilized 
with the detergent Triton X-100. Whether 
or not this toxin is associated with in- 
soluble receptors and what the rate of 
loss of toxin from this set of binding sites 
is are unknown. If this insoluble material 
is assumed not to be receptor, and the 
loss of radioactivity from these sites is 
assumed to be slow compared to degra- 
dation of toxin bound to extrajunctional 
receptors, then a half-time for degrada- 
tion is about 20 hr. If all the radioactiv- 
ity remaining in the muscle at the end of 
the second exponential (Fig. 51) were 
associated with extrajunctional recep- 
tors, then the half-time for degradation 
would be about 30 to 35 hr. Judging 
from the data on tissue cultured rat and 
chick muscle and from the data of Chang 
and Huang on degradation of toxin 
associated with rat diaphragm in vivo, 
we think a half-time of about 20 hr 
seems most reasonable. 

Among the first experiments we per- 
formed with the perfused EDL muscle 
is a test of the dependence of degradation 
on metabolic energy. In this experiment, 
illustrated in Fig. 52, the rate of degrada- 
tion was measured for many hours and 
then 10 _3 M dinitrophenol was added to 
the perfusion medium to interfere with 
ATP synthesis by the muscle. The intro- 
duction of dinitrophenol resulted in a 
tenfold decrease in the rate of loss of 
radioactivity from the muscle, confirm- 
ing that the process is an energy-requir- 
ing one. An analogous experiment with 
cultured chick muscle was reported in 
Year Book 73, p. 64. 

We have also been able to demonstrate 
that the acetylcholine receptors appear- 
ing in extrajunctional plasma membrane 




S2 8.aU 













22 days denervated 




^ * 


I 000 I 500 2000 

TIME (min) 

Fig. 51. Release of radioactivity from 22-day denervated mouse EDL muscle in organ 
culture, following brief incubation of muscle with 125 I-a-bungarotoxin. Details of culture and 
monitoring of radioactivity are in text. 

^ 9 

LU .8 




MOUSE EDL 8 days denervated 

+ DNP I0" 3 m 

1000 1500 

TIME (min) 


Fig. 52. Effect of 2, 4-dinitrophenol on the release of radioactivity from 8-day denervated 
mouse EDL, following brief incubation with 125 I-a-bungarotoxin. Measurements begin after 
the washout of most of the unbound a-bungarotoxin. 


after denervation are synthesized de until they were removed for incubation 

novo. This was accomplished using essen- with the 131 I-labeled a-bungarotoxin. 

tially the same experimental procedures The purification of these control recep- 

as were used to demonstrate the incor- tors was identical to that for the deuter- 

poration of isotopically labeled amino ated receptors. 

acids into acetylcholine receptors in cul- The receptors synthesized by EDL 
tured chick skeletal muscle. The lower muscles during maintenance in organ 
leg muscles of 10 mice were denervated culture with deuterated amino acids in 
by cutting the right sciatic nerve. Five the medium and the receptors from con- 
days later the EDL muscles were re- trol denervated muscle, now containing 
moved and maintained in organ culture, different iodinated a-bungarotoxin labels, 
During the first 3 hr in organ culture were mixed and layered on stepwise 
the culture medium contained 1 ftg/ml gradients of metrizamide-deuterium ox- 
a-bungarotoxin and all of the amino ide and centrifuged for 4 days at 33,000 
acids of the Trowell T-8 medium were rpm in a Beckman No. 40 fixed angle 
fully deuterated in the non-exchangeable rotor, which banded the receptors in a 
hydrogens. Thus, all of the existing spontaneously formed gradient. The re- 
acetylcholine receptors on the EDL mus- sultant fractionation of receptors is illus- 
cle fibers were blocked by unlabeled «- trated in Fig. 53. The receptors from the 
bungarotoxin and the newly synthesized muscles grown in organ culture with 
receptors should have incorporated the deuterated amino acids were clearly 
isotopically labeled amino acids in their more dense than those from control mus- 
polypeptide chains. Then the unlabeled cles, proving that deuterated amino acids 
a-bungarotoxin was removed by washing had been incorporated into the newly 
the muscles with many changes of me- appearing receptors in organ culture, 
dium over a 2 hr period. Then the organ Thus it is proved that acetylcholine re- 
cultures were maintained overnight in ceptors are synthesized de novo and 
culture medium, still with deuterated incorporated into the extra junctional 
amino acids. The next day the new sur- plasma membranes of denervated skele- 
face acetylcholine receptors were labeled tal muscle fibers. 

with 125 I-«-bungarotoxin by incubation Some years ago we were able to show 
for 1 hr and muscles were thoroughly that when denervated skeletal muscle 
washed to remove unbound bungaro- was maintained in organ culture dener- 
toxin. The muscles were homogenized in vation, supersensitivity developed just 
1% Triton X-100, 10 ml Tris buffer to as it does in denervated muscle in vivo, 
solubilize the labeled receptors, and the When protein synthesis was blocked by 
homogenates were centrifuged at 10,000 inclusion in the culture medium of either 
X g for 30 min to sediment insoluble puromycin or cycloheximide (both in- 
material. The supernatant, containing hibitors of protein synthesis) , the ap- 
the solubilized receptors, was concen- pearance of extrajunctional chemosensi- 
trated by pressure dialysis, layered on a tivity was blocked. Work in several 
5% to 20% sucrose gradient containing laboratories demonstrated that there is 
1% Triton X-100, and centrifuged for a large increase in the total number of 
4 hr at 62,000 rpm in a Beckman SW-65 receptors after denervation, and we have 
rotor at 20°C. The 10S peak from such shown that the receptors at the former 
gradients contained the 125 I-«-bungaro- neuromuscular junction remain at a high 
toxin labeled, deuterium-containing re- concentration there. Thus there was a 
ceptors. 131 I-«-bungarotoxin labeled con- good deal of evidence to suggest that 
trol receptors were prepared from EDL denervation supersensitivity to acetyl- 
muscles denervated for the same length choline results from the synthesis of new 
of time but maintained in the animals acetylcholine receptors. Nevertheless, it 
















< 10 

• * 

v • 

q: ' 





/ / 


1 1 







1 1 




/ i 



H 0.6 

r i 
1 i 

J «0 



7 ' 



m i 

>^ \ 


7 / 





• / 




J < 




■r / 




>• V 







1 1 1 





Fig. 53. Equilibrium density gradient sedimentation of 6-day denervated mouse EDL 
acetylcholine receptors. 131 I-a-bungarotoxin-receptor complexes from 6-day denervated muscle 
(open circles). 125 I-a-bungarotoxin-receptor complexes appearing during 15 hr in organ culture 
in medium containing 2 H-amino acids (filled circles), a-bungarotoxin-receptor complexes were 
extracted from the muscle by homogenization in 1% Triton, 10 ml Tris buffer, pH 7.8, 
and prepurified by sucrose gradient velocity sedimentation. The ratio of 125 I to 131 I values is 
shown above for peak fractions. 

was still possible that all the receptors 
involved in denervation supersensitivity 
were already present as "latent" recep- 
tors whose activation was dependent 
upon protein synthesis. This hypothesis, 
however unlikely, has now been ruled 
out. As it stands now, the appearance 
of acetylcholine receptors in extrajunc- 
tional membranes of denervated skeletal 
muscle and the metabolism of these 
receptors appear to be qualitatively like 
the corresponding phenomena in tissue- 
cultured embryonic skeletal muscle and 
even surprisingly similar in quantitative 

iodination of membrane proteins 
of Muscle Cells 

K. Tepperman 

One protein of the muscle cell mem- 
brane, the acetylcholine receptor, has 
been studied in detail by many research 
groups; and its synthesis, appearance in 
the plasma membrane, and degradation 
have been described in this laboratory. 
In order to develop a better understand- 
ing of plasma membranes, in terms of 
their structure and their metabolism, it 
would be useful to conduct similar stud- 
ies with other membrane proteins. One 



means of identifying a specific group of 
membrane proteins is to iodinate cells 
using lactoperoxidase and 125 I~. This 
method has been shown to specifically 
label proteins on the surface of the 
plasma membrane in red blood cells, L 
cells and several other cell types. Iodina- 
tion is a useful tool in describing the 
organization of the plasma membrane, 
since it can be used in studying the size 
and accessibility of surface proteins. In 
addition, the turnover of iodinated cell- 
surface proteins can be monitored by 
following the loss of iodinated proteins 
from the cell surface. In the past year, 
iodination of plasma membrane proteins 
of cultured muscle cells was carried out 
using lactoperoxidase, a peroxide gener- 
ating system, and Na 125 I. After iodina- 
tion the cells were removed from culture 
plates by several different methods. The 
proteins were solubilized in SDS buffer 
and separated by polyacrylamide gel 
electrophoresis on slab gels. Gels were 
stained to reveal the major protein bands 
and then dried and autoradiographed. 
Figures 54 and 55 show densitometer 
traces of autoradiography of 7.5% and 
15% acrylamide gels. Gels of different 
acrylamide concentrations are required 
to obtain maximum resolution of differ- 
ent size classes of polypeptides. For each 
of these gels, iodinated chick muscle cells 
in tissue culture were extracted directly 
in a buffer containing 4 M urea, 1% 
SDS and 10 mM Tris, pH 6.8. Similar 
protein distributions were obtained when 
the proteins were first precipitated with 
TCA, washed with acetone, and then ex- 
tracted in SDS buffer. Analysis of the 
different gels suggests that there are at 
least 15 clearly distinguishable iodinated 
bands, some of which may represent 
more than one polypeptide. They range 
in molecular weight from approximately 
12,000 daltons to approximately 230,000 
daltons, with the large molecular weight 
polypeptide showing the largest single 
amount of 125 I. 

In order to determine whether any of 

the iodinated bands could be nonspeci- 
fically bound serum proteins, bovine 
serum albumin was ( 125 I) iodinated by 
the chloramine T method and incubated 
with cultured muscle cells for several 
hours. After incubation cells were rinsed 
by methods used routinely after iodina- 
tion. All the 125 I-BSA could be removed, 
suggesting that nonspecifically bound 
proteins are probably not contributing to 
the protein pattern. In addition 125 I-BSA 
was used as a standard on SDS gels and 
did not coincide with any major iodinated 

Experiments were also done to test 
whether one sensitive membrane prop- 
erty, the resting potential, was changed 
as a result of iodination. Resting mem- 
brane potentials of cultured chick myo- 
tubes were monitored by Aileen Ritchie 
during the iodination reaction. At an 
iodide concentration of 10 -6 M, the con- 
centration used in the labeling experi- 
ments, the resting membrane potential 
remained at normal levels. It required 
a concentration of 1 — 4 Af iodide to sig- 
nificantly decrease resting potentials. 

Preliminary experiments have been 
done using rat muscle cells in culture for 
iodination. In these experiments, cells 
were iodinated in suspension after re- 
moval from plates by collagenase. SDS 
gel analysis of the proteins from these 
cells shows some radioactive components 
similar in molecular weight to those from 
chick cells. However, the large molecular 
band seen on autoradiographs of chick 
proteins is absent. Further experiments 
must be done to determine whether this 
difference is a species difference or a 
difference due to the preparation method. 

In some experiments, cells were pre- 
incubated with a-bungarotoxin before 
iodination. The cells were then analyzed 
both on SDS gels and on sucrose gradi- 
ents. On gels, no peak of iodination oc- 
curred at a molecular weight correspond- 
ing to bungarotoxin. On sucrose gradients, 
there was no iodinated 10S peak cor- 
responding to the position of the ACh 


receptor on either the toxin-treated or 
control plates. This indicates that neither 
the ACh receptor nor the a-bungarotoxin 
bound to it exhibit significant iodination, 
although the ACh receptor has been esti- 


mated to represent approximately 1% 
of the cell surface protein. This result 
may be important in determining the 
limits of the iodination reaction for 
identifying specific proteins. 


4 6 8 10 

Cm Migrated 




Fig. 54. Distribution of iodinated proteins on a 7.5% SDS acrylamide gel. Iodinated chick 
skeletal muscle cells were extracted in 4 M urea, 1% SDS, 10 mM Tris, pH 6.8, boiled in 
the presence of mercaptoethanol, and layered onto a Bio-Gel P-2 column to remove free 125 I. 
A portion of the excluded peak was then layered onto a stacking gel, 3% acrylamide, 0.2% 
bisacrylamide, pH 6.8, over a running gel, 7.5% acrylamide, 0.2% bisacrylamide, pH 8.8. 
Electrophoresis was carried out at 3.3v/cm for 2 hr, then at 6.6v/cm for 4 hr. Gels were 
stained, then dried and autoradiographed. The figure shows a densitometer trace of the 
autoradiogram. Also shown is a schematic drawing of the gels. Above the drawing, a measur- 
ing bar indicates the length of the stacking gel. Migration is from left to right. Positions of 
four molecular weight markers are shown: lysozyme (MW 14,800), ovalbumin (MW 43,000), 
catalase (MW 60,000) and /3-galactosidase (MW 130,000). 



15 7< 

I I 1 L 





Cm Migrated 

Fig. 55. Distribution of iodinated proteins on a 15% acrylamide gel. A fraction of the 
protein extract described in Fig. 54 was layered onto a stacking gel of 6% acrylamide, 0.2% 
bisacrylamide, pH 6.8, over a running gel of 15% acrylamide, 0.2% bisacrylamide. Gels were 
electrophoresed and prepared as in Fig. 54. 

An Investigation of the Involvement 

of Denervation in the Human 

Skeletal Muscle Diseases 

Amyotrophic Lateral Sclerosis and 

Congenital Myotonic Dystrophy 

D. M . Fambrough in collaboration with, 
D. B. Drachman* 

The muscular dystrophies constitute a 
heterogeneous group of progressive mus- 
cle disorders which are usually inherited 
and are characterized by typical clinical 
patterns of weakness. Classically, the 
muscular degeneration has been thought 

* Department of Neurology, The Johns Hop- 
kins University School of Medicine. 

to result from some primary defect in- 
herent in the muscle fiber, although such 
an abnormality has not yet been identi- 
fied in any of the dystrophies. More re- 
cently, the possibility has been raised 
that abnormalities of the motor nerves 
might be fundamental to the pathogene- 
sis of the dystrophies. This concept is 
based chiefly on similarities in the histo- 
pathology and physiological properties 
of denervated and dystrophic human 
muscle: "myopathic" histological fea- 
tures are often present in muscle of pa- 
tients with disorders of the motor nerves ; 
and fibrillations and the reduction in the 
number of functional motor units have 



been reported in a variety of neuro- 
muscular disorders, including the dystro- 

Of the group of muscular dystrophies, 
we elected to study myotonic dystrophy, 
since its features most closely resemble 
those of denervation. The muscular 
weakness is predominantly distal, as 
occurs also in peripheral neuropathies. 
Histologically, the affected muscle fibers 
show atrophy similar to that seen in 
denervation rather than destruction or 
denervation. With histochemical tech- 
niques the Type I fibers are found to be 
preferentially involved, which has been 
thought to represent a selective denerva- 
tion effect. 

The present investigation was designed 
to test the denervation hypothesis fur- 
ther by applying to human dystrophic 
muscle a widely accepted criterion of 
denervation: a high density of extra- 
junctional acetylcholine receptor sites. 
Nine patients ranging in age from 19 to 
55 years with typical myotonic dystro- 
phy gave informed consent for this study. 
All the myotonic dystrophy patients had 
family histories consistent with an auto- 
somal dominant pattern of inheritance 
and manifested a characteristic pattern 
of weakness as well as clinical electro- 
myographic and histological features of 
this disorder. As controls, two patients 
with amyotrophic lateral sclerosis (ALS) 
gave informed consent for this study. 
ALS is well known to be a disease in- 
volving progressive denervation of skele- 
tal muscle fibers. Denervated rat dia- 
phragm muscle was also used as a 

Several techniques were used to assess 
the occurrence of extra junctional acetyl- 
choline receptors. All the techniques in- 
volved the binding of 125 I-labeled a- 
bungarotoxin to acetylcholine receptors 
in thin muscle strips obtained from 
biopsy specimens. Biopsy specimens were 
dissected and incubated in medium con- 
taining 125 I-«-bungarotoxin as described 
in Year Book 72, p. 56, for biopsies of 
myasthenia gravis patients. Specimens 

were thoroughly washed to remove un- 
bound a-bungarotoxin and fixed in glu- 
taraldehyde. In some cases single muscle 
fibers were dissected from the specimens 
after staining for the enzyme acetyl- 
cholinesterase (to reveal the locations of 
neuromuscular junctions). These were 
mounted on microscope slides, and auto- 
radiographs were made. In most cases 
small bundles of muscle fiber segments 
were used in determinations of the aver- 
age density of extra junctional acetyl- 
choline receptors by a scintillation 
counting method described for dener- 
vated rat skeletal muscle in Year Book 
73, p. 61. Finally, samples of all the 
biopsies were embedded in plastic and 
1 jjm cross sections were cut and proc- 
essed for autoradiography. 

Autoradiographs of whole mounted sin- 
gle muscle fibers sites on muscle fibers of 
patients with myotonic dystrophy (after 
incubation with iodinated a-bungaro- 
toxin) revealed a normal pattern of 
bungarotoxin binding. Each fiber dis- 
played one small area of densely packed 
silver grains, reflecting the binding of a 
very large number of a-bungarotoxin 
molecules in the area of the endplate. 
The extra junctional surface of each fiber 
was virtually devoid of grains (approxi- 
mately background level). These auto- 
radiographs are interpreted to suggest 
that, qualitatively, the acetylcholine re- 
ceptors in the muscle fibers are distributed 
just as they are in the muscle fibers of 
normal, innervated adults. Autoradio- 
graphs of denervated muscle fibers pre- 
pared in the same way show a high 
density of extra junctional grains, indica- 
tive of a high level of extra junctional 
acetylcholine receptors. Most of the 
muscle fibers examined in the present 
study by whole-mount autoradiography 
were fibers of approximately normal 
diameter. However, several very small 
caliber fibers were also examined (see 
Fig. 56) , and they too appeared to be 
normal in grain distribution. 

In Table 10, values for the number of 
a-bungarotoxin binding sites per endplate 



Fig. 56. Autoradiogram of small-caliber human muscle fiber from patient with myotonic 
dystrophy, after incubation of biopsy with ^I-a-bungarotoxin. Note dense accumulation of 
silver grains at neuromuscular junction and paucity of grains over distant extrajunctional areas. 

and the density of extrajunctional sites 
on the average muscle fiber (both values 
obtained from scintillation counting) are 
presented. While there is a fair amount 
of variation in values from different 
patients, this variation may reflect nor- 
mal variability, and all values for end- 
plate a-bungarotoxin binding sites are 
above values previously reported for 
myasthenia gravis patients and similar 
to values reported for normal adults. The 
density of extrajunctional a-bungaro- 
toxin binding sites is in all cases low 
when compared with values of 500 to 
700 sites per ^m 2 obtained for 14-day 
denervated rat diaphragm fibers. The 
values for myotonic dystrophy patients 
are thus not consistent with high extra- 
junctional acetylcholine receptor density. 
The data in Table 10 indicate that the 
average muscle fiber in myotonic dystro- 

TABLE 10. Acetylcholine Receptors* in 
Muscles of Patients with Myotonic Dystrophy 


Sites per 





(mean + S.E.M.) 

(receptor/Tim 2 ) 


5.21 ± 0.83 X 10 7 



2.75 ± 0.26 X 10 7 



4.54 ± 0.41 X 10 7 



1.78 ± 0.53 X 10 7 



1.49 ±0.10 X 10 7 



2.59 ± 0.22 X 10 7 


* ^I-a-bungarotoxin binding sites. 

phy patients lacks a high density of 
extrajunctional acetylcholine receptors, 
but these data could not reveal the 
presence of a small population of muscle 
fibers with a very high density of extra- 
junctional acetylcholine receptors. There- 
fore, it was necessary to examine each 
of the muscle fibers in biopsy specimens 
to determine whether or not a dener- 
vated population of muscle fibers existed. 
For this analysis, cross sections of biop- 
sied muscle (after incubation with iodi- 
nated a-bungaro toxin) were examined 
after autoradiography. Since we could 
not be sure of the sensitivity of detection 
of denervated fibers by this technique, we 
also examined autoradiographs of de- 
nervated rat diaphragm muscle and of 
muscle biopsy specimens from two pa- 
tients with amyotrophic lateral sclerosis 
(ALS). All of the small-caliber fibers 
seen in the autoradiographs of ALS mus- 
cle (Figure 57A) and all of the dener- 
vated rat muscle fibers were overlain by 
numerous silver grains. The densities for 
denervated rat skeletal muscle and for 
ALS denervated fibers were qualitatively 
similar, suggesting that the density of 
acetylcholine receptors in atrophic ALS 
muscle fibers must be on the order of 
several hundred sites per /xm 2 . The grain 
density over atrophic fibers in areas de- 
void of large fibers seemed somewhat 
higher than the grain density over 
atrophic fibers in areas of muscle con- 
taining large numbers of large caliber 






Fig. 57. Autoradiograms of cross-sections of human muscle biopsies after incubation with 
^I-a-bungarotoxin. (A) amyotrophic lateral sclerosis (ALS) ; (B) congenital myotonic 
dystrophy. Note high grain density over small caliber fibers in ALS biopsy. 

fibers as well. This may reflect the longer 
period of denervation of the muscle fibers, 
which results in a greater accumulation 
of acetylcholine receptors in the areas 
with widespread atrophy. 

In contrast to the ALS and denervated 
rat muscle, the muscles of myotonic 
dystrophy patients did not contain any 
fibers which displayed large numbers of 
extra junctional a-bungaro toxin binding 
sites (Fig. 57B). We examined several 
hundred small caliber fibers as well as 
about ten times that many larger caliber 
fibers in autoradiograms of muscles from 
patients. Also, grain counts were made 
of the perimeter area of 137 small caliber 
fibers and 86 large caliber fibers. The 
average grain densities were very similar 
and the absolute grain densities were 
so low that statistical analysis of indi- 

vidual fibers was not meaningful. The 
grain density for large caliber fibers was 
0.73 per 100 /mi 2 and that for the small 
caliber fibers was 0.82 per 100 /xm 2 . These 
densities approximate the density of 
grains that occurred as background on 
these slides. In contrast to these negative 
data for myotonic dystrophy patients, 
grain counts for a biopsy of one ALS 
patient, done together with the congeni- 
tal myotonic dystrophy material, re- 
vealed a grain density over atrophic 
fibers which was 3 per 100 /xm 2 higher 
than the grain density over the large, 
presumably still innervated fibers. 

The results of this study demonstrate 
that there is no population of skeletal 
muscle fibers in myotonic dystrophy 
muscles which displays a high extra- 
junctional acetylcholine receptor density. 


Since an increase in extrajunctional re- is perhaps the strongest argument against 

ceptors has been found to be an invari- the denervation hypothesis. Another pos- 

able consequence of denervation of mam- sibility is that spontaneous muscle ac- 

malian muscle fibers, our findings suggest tivity prevents a denervation-like in- 

that there is not a significant population crease in extrajunctional receptors in 

of denervated muscle fibers in myotonic denervated muscle fibers of myotonic 

dystrophy. dystrophy patients. It has been shown 

These data do not prove that dener- experimentally that electrical stimula- 

vation is not involved in mytonic dystro- tion of denervated skeletal muscle can 

phy. One possibility is that some fibers largely prevent or reverse the develop- 

are denervated for only a short time and ment of extrajunctional acetylcholine 

are quickly re-innervated by neighbor- sensitivity. Could the spontaneous ac- 

ing motor nerves. If this were the case, tivity of myotonic muscles have a similar 

the small fibers should nevertheless be effect? Although the amount of activity 

the "denervated" ones in transition be- of myotonic muscles has not been quanti- 

tween one motor unit and another and tatively measured, it would seem to be 

should therefore show an increase in far less than the regular, continual stimu- 

acetylcholine receptors. The absence of lation required to prevent the denerva- 

such an increase in the atrophic fibers tion effect. 


K. J. Mutter, in collaboration with Uel J. McMahan* 

The central nervous system of the physiologically and shown to mediate 
leech Hirudo medicinalis provides a fav- reflexive behavior. It has also been pos- 
orable preparation for studying a variety sible experimentally to make permanent 
of fundamental neurobiological problems, changes in the synaptic interactions be- 
particularly those dealing with synaptic tween certain neurons. In order to corn- 
transmission and integration. Its organi- plement these physiological studies, over 
zation is so simple and reliable that a the past year we have examined in some 
single neuron impaled with microelec- detail the structure of specific neurons 
trodes in one animal is found to perform with the aim of locating and character- 
like a corresponding cell in another. The izing their synapses, 
central nervous system of the leech con- The synapses of cells in the ganglion 
sists of a chain of identical segmental are between neuronal processes that 
ganglia; each ganglion controls the form an intricate meshwork, the neuropil, 
stereotypic patterns of behavior of its in the center of the ganglion. This ar- 
body segment. Of the 350 neurons in rangement has made it difficult to trace 
each ganglion, the functions of about the origins of particular profiles seen in 
100 have been identified. Most of the the electron microscope and even to fol- 
sensory and motor neurons can be recog- low the continuity of selectively stained 
nized from animal to animal by the processes in the light microscope. We 
relatively constant shapes and positions have therefore injected horseradish per- 
of their cell bodies. Synaptic contacts oxidase through a recording micropipette 
between some mechanosensory and into specific neurons to mark their finest 
motor neurons have been characterized processes. Peroxidase forms a dense re- 
action product that is visible in the 

* Department of Neurobiology, Harvard injected neurons' cytoplasm throughout 

Medical School. the ganglion; ultrastructural details are 



fl « GQ 

O J3 -p 
■ .H -+J o 

c6 O o 
hfi s-, o 

2 «-S 


oa "■' co 


O co r> 

2 ° 

Q< co 


CD +? 

d '53 




CD ^3 

s -£ -^ 
^ ° z 

-+-J CO 

■r 1 IB In 


T3 CO p 

CD •— i 

■P (DO 


CO . -+J 

c3 CD ,-Q 

> f-h bD 

^ O ■F 

CO OJ fl 

fi "S ° 

■ O CD 

-£ S bJD 

O M s_ 
CD C3 

CD i « 
£ .SI 

fl « o 

c3 -d 

ri a co 

(Da fl 
S3 t3 3 

S? CD fe 
O TJ co 

CO -^ 

Si £ CD 
cu O ^ 

CD o J-; 
*-• O & 
co T3 Sh 

S b^ 

a;g O 

03 CD g 

„ 'tS co 

CD N ^ 

«J rt c 

co B 

® S CD 

2^ * 

Pm ,-g M 


t-H P 

c3 <3 

g a 



Fig. 59. A cell as in Fig. 58 reacted with diaminobenzidine tetrahydrochloride and prepared 
for electron microscopy. Synaptic vesicles are visible in the injected processes, and at the 
lower left the sensory neuron is making synaptic contact with two unidentified processes. 



preserved so that synaptic components 
can be recognized electron-microscopi- 
cally and correlated with structures seen 
in the light microscope. We have found 
that each type of mechanosensory or 
motor neuron has a distinctive distribu- 
tion of synapses and processes within 
the ganglion. Figure 58, for example, 
shows light micrographs from the whole 
mount of an injected pressure sensory 
neuron as its main process passes through 
the neuropil. Fine branches emerge from 
the main trunk; at their tips and along 
their length are clusters of fingerlike 
processes that are in sites of dozens of 
synapses, one of which is illustrated in 
the electron micrograph in Fig. 59. 
Prominent swellings also occur along the 
fine processes of other sensory neurons. 
These varicosities are the sites of con- 

centrations of vesicles; the arrangement 
of membrane specializations indicates 
that they are presynaptic "terminals. " 
Some of the vesicle-containing varicosi- 
ties are also postsynaptic. Motoneuron 
processes are studded with spines a few 
microns long and a fraction of a micron 
in diameter. The motoneurons do not 
contain prominent clusters of vesicles 
but receive extensive synaptic contact 
by varicosities, chiefly upon their spines 
and small branches. 

Armed with a detailed description of 
identified neurons in the normal animal, 
it should be possible to trace the regen- 
eration of connections and to detect any 
morphological basis for physiological 
and behavioral changes that can be pro- 
duced by experimental manipulation of 
the leech and its nervous system. 




K. Ozato, W. H. Adler* J. D. Ebert, and L. Huang 
assisted by Delores Somerville and Bessie Smith 

Regulation of Transplantation 

Immunity by T Lymphocytes : 

Effects of Mitogens 

K. Ozato, J. D. Ebert, and W. H. Adler* 

We have continued to explore the 
mechanism of T cell recognition. Since 
the recognition of alloantigens is achieved 
by cell to cell interaction, which is diffi- 
cult to analyze directly, we have used 
simple probes to approach the problem, 
that is, we have modified the immune 
response by the mitogens Concanavalin 
A (ConA) and bacterial lipopolysaccha- 
ride (LPS). Previously we had estab- 
lished an experimental model system in 
which the proliferative response (mixed 
lymphocyte reaction (MLR) ) and the 
generation of cytotoxic lymphocytes 
(GCL) could be assayed. To be confi- 

* Gerontology Research Center, National In- 
stitutes of Health, Baltimore, Maryland. 

dent that we were dealing with T cells, 
we used cortisone-resistant thymocytes 
as the responding cells, thereby exclud- 
ing contamination by B cells or their 
product, anti H-2 antibody, which are 
known to interfere with T cell mediated 
immunity. We focused on appropriate 
combinations of strains in which the 
gene products other than H-2 antigens 
were avoided; thus we chose the BIO vs. 
B10.D2 system (H-2 b vs. H-2 d ). We 
found that the B cell mitogen LPS 
(added at the initiation of culture) en- 
hanced the generation of cytotoxic lym- 
phocytes when the responding cells were 
stimulated with suboptimal doses of allo- 
antigens (Fig. 60). On the other hand, 
mitogenic concentrations of ConA (5 and 
10 /*g/ml) suppressed cytotoxic activity 
regardless of alloantigen dose. These 
effects were found to affect the early 
events of lymphocyte differentiation; the 
addition of mitogens 24 hr after the initi- 















" Con A 

ly 5y lOy 50y 




ly 5y lOy 50y 3y 

I :o.3 

Responding Cells: Stimulating Cells 
BIO (H-2 b ) : BI0.D2 (H-2 d ) 

>— ConA-J LPS 


!— ConA-l LPS 

ly 5y lOy 3y 

ly 5y lOy 3y 



Fig. 60. Effect of ConA or LPS on the generation of cytotoxic lymphocytes (BIO vs. B10.D2). 
1 X 10 7 cortisone-resistant thymocytes were sensitized with B10.D2 spleen cells at different 
cell ratios (expressed as responding vs. stimulating cells). ConA at the concentrations indi- 
cated or LPS (3 Atg/ml) was added at the initiation of culture. Cytotoxic activity per chamber 
on day 5 is expressed as percent 51 Cr-release from target cells (P815). Each column represents 
the mean of duplicates ± standard deviation. 


1=0.5 h 0.25 

Responding Cells (5xl0 6 /ml) 
St i mula ting Cells 

Fig. 61. Effect of LPS on the proliferative response on day 3 to allogeneic stimulation. 
1.25 X 10 6 BIO CRT were cultured with B10.D2 spleen cells at different cell ratios. LPS 
was added at the initiation of the cultures (filled triangles) or at 24 hr thereafter (open 
triangles). Controls are indicated by filled circles. Cells were labeled with 3 H-TdR (1 /xCi/ml) 
for 18 hr at different time-points. Each value represents the mean of triplicates. 



ation of culture induced neither enhance- 
ment nor suppression. Selective enhance- 
ment of the GCL by LPS in lower doses 
of alloantigens was observed in the 
MLR, too (Fig. 61). The response was 
about twice as high. Taking into account 
the previous rinding that LPS lowers the 
mitogenic doses of ConA in T cells, LPS 
appears to potentiate T lymphocytes 
when suboptimal doses of stimulatory 
alloantigens are given. 

In this work considerable effort was 
devoted to exploring the mechanism of 
suppression, especially to determining 
the possible involvement of suppressor 
T cells. Recently it has been shown that 
both antibody response and T cell-medi- 
ated immunity are regulated, so as not 
to "overshoot" the response, by the lym- 
phoid population itself, in most cases by 
T cells, which have been termed "sup- 
pressor T cells." Since the suppressor 
cell concept could give a clue to the 
nature of immunological tolerance, we 
asked whether suppressor T cells might 
play a role in our system. ConA-acti- 
vated thymocytes were tested for their 
ability to suppress the generation of 
cytotoxic activity comparably with solu- 
ble ConA. We tested three different lym- 
phocytes with regard to suppressor ac- 
tivity. The results are shown in Table 
11. Spleen cells or lymph node cells 

TABLE 11. Effect of ConA-Activated Cells on 
the Generation of Cytotoxic Lymphocytes* 


L Cr-Release 

Control 89.3 ± 7 

ConA-activated thymocytes added 90.1 ± 4 

ConA-activated spleen cells added 26.0 ± 2 
ConA-activated lymph node cells 

added 23.8 ± 3 

* 1 X 10 7 BIO cortisone-resistant thymocytes 
were sensitized with 5 X 10 8 B10.D2 spleen cells 
treated with mitomycin C. ConA-activated 
lymphocytes (1 X 10 6 ) obtained 48 hr after in- 
cubation were added to the culture. Cytotoxic 
activity per chamber on day 5 is expressed as 
the percent Cr 51 release from P815 target cells 
(mean of duplicate ± SD). 

stimulated by ConA could significantly 
suppress the GCL when they were added 
at a ratio of 1/10. Suppression was de- 
pendent on the presence of T cells be- 
cause anti-0 treatment in the presence of 
complement completely abolished the 
activity. It was independent of cell divi- 
sion, however, for mitomycin C treat- 
ment did not impair the activity. Most 
remarkably, however, thymocytes stim- 
ulated by ConA did not suppress the 
immune response. Thus we can draw two 
conclusions: First, although transplanta- 
tion immunity can be regulated by sup- 
pressor T cells, not all T lymphocytes 
function as suppressors: Thymocytes 
were inactive. Second, the mechanism of 
suppression observed in our experimental 
system is not attributable to suppressor 
T cells, since we have used only thymo- 
cytes as the responding cells. Therefore, 
the modifications of the lymphocyte sur- 
face by ConA are assumed to be the 
immediate cause. 

Modification of the Lymphocyte 
Surface by ConA 

K. Ozato and J . D. Ebert 

The finding that suppressor T cells are 
not responsible for the suppression just 
described led us to explore further the 
possibility that the interaction of the 
ConA molecule with the lymphocyte 
surface can modify the subsequent im- 
mune response. 

Therefore, further modifications of 
lymphocyte surfaces were carried out 
both on responding lymphocytes (corti- 
sone resistant thymocytes) and on stimu- 
lating lymphocytes (mitomycin C- 
treated allogeneic or syngeneic spleen 
cells), by "coating" lymphocytes with 
native (tetravalent) ConA and succinyl 
(divalent) ConA. Coating of the cells 
with ConA alone did not induce a mito- 
genic response. We discovered that when 
the responding cells were coated with 
ConA, the proliferative response to allo- 
geneic cells was enhanced 5 to 7 times. 
A comparable enhancement was noted 



even when stimulating cells were coated a highly significant proliferation in syn- 

with ConA (Fig. 62). Perhaps the most geneic combinations. This response was 

striking observation was that lympho- observed by coating either responding or 

cytes treated with ConA could manifest stimulating cells (Table 12). Native 

• 7 




Responding CRT: Pretreated with Con A 


/ \ 


- o— ON-Con A 
A— A S-Con A 

- o—oMito- treated CRT 


Stimulating Spleen Cells: Pretreated with Con A 





_— -4 -\ 

No Spleen 1:0.15 1:0.25 1:0.5 I: I No CRT hO.15 10.25 1:0.5 1:1 

Cells Responding Cells (BIO) : Stimulating Cells (BI0.D2) 

Fig. 62. Left: Proliferative response by ConA-pretreated cortisone-resistant thymocytes. 
1.7 X 10° normal BIO cortisone-resistant thymocytes or BIO cortisone-resistant thymocytes 
pretreated with 50 /^g/ml of native ConA or succinyl ConA were sensitized with different 
numbers of allogeneic spleen cells (B10.D2), expressed as the ratio of responding to stimulating 
cells. 3 H-TdR (1 /iCi/ml) was added 52 hr after the initiation of culture. Cells were labeled 
for 18 hr. Cortisone-resistant thymocytes pretreated with native ConA followed by subsequent 
treatment of mitomycin C were also tested. Each value indicates the average of triplicate 
culture ± standard deviation (SD). Right: Proliferative response by cortisone-resistant 
thymocytes sensitized with spleen cells coated with ConA in the one-way MLC. 1.7 X 10 e 
normal BIO cortisone-resistant thymocytes were sensitized with varying numbers of normal 
B10.D2 spleen cells or spleen cells pretreated with 50 £ig/ml of native ConA or succinyl 
ConA. Cortisone-resistant thymocytes treated with mitomycin C were also sensitized with 
native ConA pretreated spleen cells. 3 H-TdR (1 juCi/ml) was added 52 hr after the initiation 
of culture. Cells were labeled for 18 hr. Each value indicates the average of triplicate culture 
± SD. 

TABLE 12. Effect of ConA Coating of Lymphocytes on Triggering in Syngeneic Combination 

Responding Cells 
Pretreated with ConA* 

3 H-TdR 

Stimulating Cells 3 H-TdR 

Pretreated with ConAt Incorporation 

Native ConA 
Succinyl ConA 

197 ± 10 

21272 ± 1018 

1613 ± 502 

native ConA 
succinyl ConA 

94 ± 4 

20393 ± 1205 

2790 ± 212 

* Responding cortisone-resistant thymocytes (B10.D2) were treated with 50 yug/ml of ConA 
for 30 min, after which they were washed, resuspended in fresh medium, mixed with syngeneic 
spleen cells, and treated with mitomycin C. 

f Stimulating spleen cells (B10.D2) were pretreated with 50 Mg/ml of ConA for 30 min, 
after which they were treated with mitomycin C. They were then mixed with cortisone- 
resistant thymocytes. 3 H-TdR (1 /mCi/ml) was added 52 hr after culture. The values represent 
mean of duplicate ± SD. 




























5-Day Responses 


I : 0.35 
I •• 0.2 
I •■ 0.15 

2 10 

ConA Concentration (/x.g/ml) 


Fig. 63. GCL by cortisone-resistant thymocytes pretreated with ConA. 1 X 10 7 BIO. 
Cortisone-resistant thymocytes treated with varying concentrations of native ConA (solid 
lines) or succinyl ConA (dotted lines) were sensitized with different numbers of B10.D2 spleen 
cells (expressed as the ratio of responding to stimulating cells). Cytotoxic activity per cham- 
ber on day 5 is expressed as percent of 51 Cr-release from target P815 cells (mean of duplicate 

ConA was much more effective than 
succinyl-ConA. We confirmed that when 
spleen cells are used as stimulators, all 
the syngeneic combinations tested (H-2 k , 
H-2 b , H-2 d ) showed the response. Al- 
though syngeneic cells have not been 
considered to be stimulatory, our findings 
show that under those conditions, T cells 
may be stimulated in the syngeneic com- 
bination. We look to a further study of 
this syngeneic response for possible clues 
to the understanding of self-recognition. 
The consequences of membrane modi- 

fication were studied further. The capa- 
bility of GCL was completely suppressed 
when either responding or stimulating 
cells were modified by ConA. The severe 
suppression seemed to be related some- 
how to the accelerated proliferative 
response, since only at doses where 
proliferation was enhanced was the sup- 
pression of GCL demonstrated (Fig. 63). 
Thus it is clear that the immune response 
of the lymphocytes whose cell surface 
has been modified is regulated in both 
the proliferative and effector phases. 



The Study of ConA Binding Sites on 
Murine Thymic Lymphocytes 

K. Ozato and J. D. Ebert 

Last year we reported (Year Book 73, 
p. 81) the nature of the thymocyte re- 
sponse to the T cell-specific mitogen, 
ConA. We dealt especially with the 
synergistic effect of a B cell-specific 
mitogen, bacterial lipopolysaccharide 
(LPS). Thereafter, we studied the bind- 
ing of mitogens to lymphocytes to try 
to delineate the relationship between cell 
surface events and subsequent DNA syn- 
thesis. First native ConA was labeled 
with tritium by acetylation to visualize 
the binding. We have now compared two 
groups of thymocytes — cortisone-resist- 
ant thymocytes and normal thymocytes 
that are different in their mitogenic re- 
sponsiveness, the former being about 
seven times more vigorous than the latter 

in mitogenic responsiveness to ConA. 
Our results show that these two lympho- 
cytes have an almost identical binding 
capacity for ConA at both 0°C and 37°C, 
that is, they have the same binding sites 
with the same kinetic properties (Fig. 
64). According to our calculations, there 
are about 7 X 10 5 receptors on one 
thymocyte. The number of receptors per 
thymocyte appeared to be far greater 
than that required for mitogenic stimu- 
lation, for the saturation dose for bind- 
ing was 600-fold higher than the mito- 
genic dose (1 fig/m\). We further investi- 
gated whether ConA binding could be 
inhibited or otherwise modified by other 
mitogens such as LPS or phytohemag- 
glutinin-p (PHA-P). In the responsive- 
ness to ConA tested in the thymocyte 
system, LPS showed a cooperative role, 
while PHA-P was inhibitory. Thus, these 
two reagents were considered useful 











20 30 40 

Minutes After Incubation 



Fig. 64. Kinetics of 3 H-ConA binding to thymocytes. Thymocytes (1 X 10" cells in 0.2 ml) 
were labeled with 50 /*g/ml of 3 H-ConA for various periods at 37 °C (open triangles, cortisone 
resistant; open circles, normal) or at 0°C (filled triangles, cortisone resistant; filled circles, 
normal) . Each value represents the average of triplicates. 



— 100 



2 50 



5 25 50 

PHA ( ^g/ml) 

250 500 

0.2 2 10 

LPS ( fxq /ml 


Fig. 65. Effect of LPS or PHA-P pretreat- 
ment on the binding of 3 H-ConA to thymo- 
cytes. Thymocytes were pretreated with various 
concentrations of LPS for 30 min, then labeled 
with 50 Mg/nil (filled triangles) or 100 /wg/ml 
(open triangles) of 3 H-ConA. In another series 
thymocytes were pretreated with various con- 
centrations of PHA-P before they were labeled 
with 50 /ug/ml (filled circles) or 100 Aig/ml 
(open circles) of 3 H-ConA. 

probes for exploring the binding of mito- 
gens in relation to subsequent triggering 
of proliferation. As shown in Fig, 65, we 
found that when thymocytes were pre- 
treated with PHA-P, the subsequent 
binding to ConA was significantly re- 
duced, whereas LPS did not change the 
binding property. Thus, the inhibition of 
mitogenic response to ConA by PHA-P 
might be attributable to the hampering 
of ConA binding to thymocytes. 

Enhancement of Surface Permeability 
of Thymocytes by ConA 

K. Ozato, L. Huang, and J. D. Ebert 

The binding of mitogens to lympho- 
cytes provides an attractive experimental 
model system to investigate signals from 
the cell membrane for the triggering of 
DNA synthesis. The influx of Ca+ 2 has 
been proposed as a secondary message 

for signaling the onset of proliferation 
(Luckasen et al., 1974). To obtain more 
detailed information about the uptake 
of Ca+ 2 , we studied the relations between 
the binding of mitogens to thymocytes 
and the transport of substances across 
the cell membrane. 

As shown in Fig. 66, the T-cell mito- 
gens PHA and ConA enhanced 45 Ca 
uptake almost equally in cortisone- 
resistant and normal thymocytes. The 
B-cell mitogen, LPS, did not change 
45 Ca uptake. The increase in 45 Ca up- 
take, however, did not correspond to the 
dose-response relationship of ConA mito- 
genicity. No detectable enhancement 
was found at mitogenic concentrations 
(1 ju,g/ml). Significant enhancement was 
demonstrated only at higher concentra- 
tions. The uptake of 45 Ca increased lin- 
early at least for 60 min of incubation 
(Fig. 67) and was entirely dependent on 
the presence of ConA, since the incre- 
ment of uptake ceased immediately after 
the specific ConA inhibitor, «-MM, was 
added. Interestingly, ConA not only 
affects Ca+ 2 uptake, but also enhances 
the uptake of Cr0 4 ~ in thymocytes (Fig. 
68). The enhancement was again ob- 
served at ConA concentrations higher 
than mitogenic doses. Moreover, we dis- 
covered that the binding of ConA to 
thymocytes changes permeability of 
trypan blue, a dye with a molecular 
weight of about 960 which is widely used 
for viability testing. After 40 min incu- 
bation of cells in the presence of 50 
/xg/ml of ConA, approximately 20% to 
40% of the cells were stained (Table 
13). These stained cells were not dead 
but merely reflecting a transient increase 
in surface permeability. This conclusion 
was based on the following observation. 
Thirty minutes after incubation when a 
significant fraction of the cell population 
became stainable, «-MM was added to 
the incubation mixture, and the dye ex- 
clusion test was performed 10 and 30 
min later (Table 13). By this treatment, 
most of the cells were no longer stain- 
able. Strong agglutination by ConA was 












- 100 




Con A 





/ O 





▲ J 

f A 


y lps 





I 1 



10 50 100 

Concentra t ion 

200 300 

of M itogens ( /*g / ml ) 


Fig. 66. Effect of mitogens on 45 Ca uptake by thymocytes. 5 X 10 a normal thymocytes 
(open symbols) or cortisone-resistant thymocytes (filled symbols) (both from CBA strain) 
in 0.2 ml RPMI 1640 containing various concentrations of mitogens were incubated in the 
presence of 1 ^Ci 45 Ca for 40 min. 

also reversed. To test whether this per- 
meability change induced by ConA is 
also valid for larger molecules, we have 
examined the uptake of 14 C-inulin which 
has a molecular weight of 5000 to 7000. 
No detectable increase in inulin uptake 
was observed at ConA concentrations up 
to 100 jug/ml. Thus, the size of the 
"pores" on the thymocyte surface in- 
duced by ConA binding is likely to fall 
between the sizes of trypan blue and 

In conclusion, we believe it is prema- 
ture to hypothesize that Ca+ 2 influx is 
directly responsible for the lymphocyte 
activation by mitogens in the system we 
have studied because increase in Ca+ 2 
uptake does not occur at mitogenic doses 
and because the uptake of other mole- 
cules and ions is also stimulated by 
mitogen binding. 

TABLE 13. Effect of ConA on Dye Exclusion 
by Thymocytes 

Stained Cells (%)* 



40 60 




min min 




2.7 2.3 

ConA (10 /ig/ml) 



23.4 31.6 

tConA (10 /xg/ml) 

+ MM 

3.4 3.0 

ConA (50 /ig/ml) 



38.6 48.1 

tConA (50 fig/ml) 

+ MM 

2.9 3.7 

* 5 X 10 6 normal thymocytes in 0.2 ml of 
RPMI 1640 in the presence of ConA were incu- 
bated at 37°C. 0.1 ml trypan blue (0.5% in 
PBS) was added at the time-points indicated. 
More than 200 cells were counted for each sam- 
ple. Each value represents mean of duplicate. 

fa-MM (10 _1 M) was added to the cell 
suspension in the presence of ConA 30 min 
after the incubation. Dye exclusion test was 
performed at the time-points indicated. 









2 50 - 

20 30 40 50 

n cuba Hon Time (min.) 


Fig. 67. Kinetics of 45 Ca uptake by thymocytes: Effect of inhibitors. 5 X 10 6 normal 
thymocytes were incubated in 0.2 ml RPMI 1640 containing 1 /iCi 45 Ca in the presence 
(filled triangles) or absence (open circles) of ConA (50 /*g/ml). a-MM (10 _1 M) (open 

triangles) or galactose (10 _1 M) (X X) was added 30 min after incubation (upper arrow) 

to ConA containing cell suspensions. ConA (50 /tg/ml, open circles) was added to control 
cell suspensions 30 min after incubation (lower arrow). 



20 30 40 50 

Incubation Time (Minutes) 

Fig. 68. Kinetics of 51 CrO~4 uptake by thymocytes. 5 X 10 e normal thymocytes were 
incubated in 0.2 ml RPMI 1640 containing 1 /iCi Cr 51 in the presence of various concentrations 
of Con A. 


R. O'Rahilly and E. Gardner 

The Carnegie Embryological Collec- 
tion (including the Bluntschli material), 
which was moved to Davis in July 1973, 
has now been installed in "Carnegie 
Village," a group of half a dozen build- 
ings at the University of California, in 
a delightful setting halfway between the 
School of Medicine and the California 
Primate Research Center. The transfer 
was made on 11 June 1974, and a formal 
opening is being planned for November 
1975. Several visitors have studied the 
Collection during the past year, includ- 
ing Arthur T. Hertig, who, with John 
Rock, contributed so many of the early 
human embryos to the Collection. 

Forty-nine specimens, more than a 
dozen of which belong to the embryonic 
period, were acquired during the year. 

All inquiries concerning the Collection 

should be addressed to Professor R. 
O'Rahilly or Professor E. Gardner, Car- 
negie Laboratories of Embryology, Uni- 
versity of California, Davis, California 

Developmental Stages in 
Human Embryos 

Work continues on the revision of 
stages 10 to 23 and on the tabulation of 
the timing and sequence of develop- 
mental events in staged human embryos. 
A detailed account of the development 
of the limbs by O'Rahilly and Gardner 
has been submitted for publication. Pre- 
liminary work by Alexander Barry has 
been begun on the establishment of a 
computer catalogue of all the prenatal 
mammalian material. 



An interesting sequel to Blechschmidt's 
Die pranatalen Organsysteme des Men- 
schen (Hippokrates Verlag, Stuttgart, 
1973), in which the Carnegie system of 
staging was adopted, is Gasser's Atlas of 
Human Embryos (Harper and Row, 
Hagerstown, Md., 1975), in which pho- 
tomicrographs in some 15 staged em- 
bryos (14 of which belong to the Car- 
negie Collection) are reproduced. A 
number of drawings are also included. 
In stage 23 (8 postovulatory weeks), for 
example, 38 photomicrographs of Car- 
negie embryo No. 4570 (30.7 mm C.-R.) 
are shown with detailed labels and draw- 
ings to indicate the levels of the sections. 

Development of the Nervous System 

Work continues on the "Developmental 
Neurobiology of Primates," for which a 
research grant was awarded to Ernest 
Gardner, as mentioned in last year's 
Report. A chapter on "The develop- 
mental anatomy and histology of the 
human central nervous system" is being 
prepared by O'Rahilly and Gardner for 
the Handbook of Clinical Neurology, 
edited by P. J. Vinken and G. W. Bruyn 
(North Holland Publishing Co., Amster- 

dam) . Volumes 26 and 27 on "Congenital 
Malformations of the Brain and Skull" 
are being edited by N. C. Myriantho- 
poulos. An article on "The Dandy- 
Walker and Arnold-Chiari malforma- 
tions" by Gardner, O'Rahilly, and Prolo 
is due to appear in the Archives of 
Neurology. A paper on "The prenatal 
development of the human eye" by 
O'Rahilly has been accepted for publi- 
cation in Experimental Eye Research. 
An account of "The nerve supply and 
conducting system of the human heart 
at the end of the embryonic period 
proper" has been submitted for publica- 

Witold Wozniak, Medical Academy of 
Poznan, Poland, spent a month in Davis 
studying the development of the human 
nervous system. He prepared projection 
drawings and photomicrographs of the 
brain at stage 21 (7% postovulatory 
weeks) , and it will be of interest to com- 
pare in detail the level of development 
with that at stage 23 (8 weeks), already 
under investigation by Gardner and 
O'Rahilly. Wozniak is also engaged in a 
study of the developing human spinal 
cord with the aid of both light and elec- 
tron microscopy. 


The dissemination of research findings, 
through presentations at conferences and 
meetings of learned societies, publica- 
tions, and teaching serves several func- 
tions: It provides an opportunity for 
criticism and exchange, leading ulti- 
mately to verification of discovery, and 
advance — the word "ultimately" being 
crucial, for verification may be preceded 
by failure, reexamination and redirection. 

Teaching promotes another crucial 
role: it enables our staff to attract new 
investigators of high quality by creating 
an awareness of the forces impinging 
upon and fostering innovation. Here our 
concern is not so much with increasing 
the number of developmental biologists 
as it is with providing enrichment and 

rare opportunities for intellectual growth 
for gifted young men and women by the 
preceptorial approach, long the hallmark 
of enlightened postdoctoral education. 

International conferences and sym- 
posia in which members of the Depart- 
ment participated during the year 
included the following: The Fifth Con- 
ference, "De la Physique Theorique a 
la Biologie (Versailles, France) ; a sym- 
posium on Mechanism of Synaptic Ac- 
tion, in connection with the International 
Congress of Physiological Sciences 
(Jerusalem, Israel) ; a Conference on 
Ribosomes (St. Johns, Newfoundland) ; 
a Conference on Tests of Teratogenicity 
in vitro (Woods Hole, Massachusetts) ; 
a Symposium on Cancer Biology at the 


Given Institute of Pathobiology (Aspen, Yvette, France; the Cold Spring Harbor 
Colorado); the Conference on Recom- Laboratory; Institute for Cancer Re- 
binant DNA Molecules (Asilomar, Cali- search; Marine Biological Laboratory; 
fornia) ; the Ninth Leukocyte Culture M. D. Anderson Hospital; National In- 
Conference (Williamsburg, Virginia) ; stitutes of Health; National Jewish Hos- 
and a Conference on Selected Aspects of pital; and Weizmann Institute. 
Developmental Genetics (Iowa City, Members of the Department took part 
Iowa). in meetings of learned societies, includ- 
Among the other conferences engaging ing, in addition to those already men- 
the interest of members of the group, the tioned, the American Association for the 
following should be mentioned: The Uni- Advancement of Science, American Asso- 
versity of Minnesota Symposium on the ciation of Anatomists, American Society 
Interface between Molecular Biology of Biological Chemists, American So- 
and -Medicine (Minneapolis) ; the New ciety for Cell Biology, American Philo- 
York Academy of Sciences' Conference sophical Society, American Physiological 
on Human Uses of Molecular Genetics; Society, American Society for Micro- 
Five Gordon Research Conferences, biology, Biophysical Society, Federation 
dealing with Biological Regulatory of American Societies for Experimental 
Mechanisms, Chromatin, Developmental Biology, National Academy of Sciences, 
Biology, Lipid Metabolism, and Nucleic Society for Developmental Biology, So- 
Acids; the UCLA-ICN Winter Sympo- ciety of General Physiologists, and So- 
sium on DNA Replication (Squaw Val- ciety for Neuroscience. 
ley, California) ; and a Symposium on Advisory and consultative services in- 
Excitable Cells in Tissue Culture (At- eluded membership on the editorial 
lantic City). boards of Anales del Desarrollo, Cell, 
Lectures were presented on a number Developmental Biology, Journal of Em- 
of campuses, including Albert Einstein bryology and Experimental Zoology, Ex- 
College of Medicine; Case Western Re- cerpta Medica (section on Human De- 
serve University; Colorado State Uni- velopmental Biology), Current Topics in 
versity; Columbia University; Duke Developmental Biology, Oceanus, and 
University School of Medicine; Emory Quarterly Review of Biology. 
University; Erindale College; Hebrew Members of the staff also acted in 
University, Jerusalem; Loyola College; these capacities: Chairman, Assembly of 
McGill University; Moravian College; Life Sciences, National Academy of 
Princeton University; Saint Louis Uni- Sciences-National Research Council; 
versity; Stanford University; State Uni- Member of the Board of Governing 
versity of New York at Buffalo ; Temple Trustees and the Board of Scientific 
University; the Universities of California Overseers, Jackson Laboratory; Trustee, 
at Berkeley and Santa Barbara; Chi- President, and Director, Marine Biologi- 
cago, Cincinnati, Colorado, Maryland, cal Laboratory ; Member of the Corpora- 
Missouri, North Carolina, Paris, Penn- tion, Woods Hole Oceanographic Institu- 
sylvania, Rochester, Tennessee, Texas, tion; and Trustee of International 
Virginia, Wisconsin and Yale Universi- Medical Congress, Ltd. 
ties. Other posts occupied by members of 
Special presentations included The the Department include the following: 
Jessup Lectures at Columbia University in the American Institute of Biological 
and a series of lectures as Regents' Lee- Sciences, member, Council of Past Presi- 
turer at the University of California. dents; in the Institut de la Vie, Chair- 
Members of the group also spoke at man, World Committee on the Formative 
hospitals and research centers, among Weeks of Human Life; in the National 
them the CNRS Laboratory, Gif-sur- Academy of Sciences, member, Commit- 



tee on Science and Public Policy; mem- 
ber, Coordinating Committee for Air 
Quality Studies; and member, Commit- 
tee on Energy ; in the National Institutes 
of Health, member, Board of Scientific 
Counselors, National Institute of Child 
Health and Human Development, mem- 
ber, Gerontology Research Center Re- 
sources Advisory Committee, member, 
Physiological Chemistry Study Section; 
and member-elect, Cell Biology Study 
Section; and in the Society for Develop- 
mental Biology, President. 

Members of the Department served on 
several visiting committees, including 
those for the Departments of Biology at 
Brookhaven National Laboratory, Har- 
vard University, Massachusetts Institute 
of Technology, and Princeton Univer- 
sity, and the External Advisory Commit- 
tee of the University of Chicago's Cancer Center. 

Staff members taught in the EMBO 
course on Molecular Genetics of Higher 
Organisms at Port-Cros, France, and in 
courses at the Marine Biological Labora- 
tory, as well as at Johns Hopkins (in 
the Departments of Biology, Biophysics 
and Pediatrics). 


The roster of speakers at the seminars 
organized by the Department to serve 
all those working in developmental biol- 
ogy in the region included: Bruce Alberts 
(Princeton University) ; E. X. Albu- 
querque (University of Maryland) ; T. 
Blumenthal (Indiana University) ; 
Charles Boder (Geneva) ; Diana Card 
(University of California, Los Angeles) ; 
David Clayton (Stanford University) ; 
Marco Crippa (Geneva) ; James Darnell 
(Rockefeller University) ; Laura de 
Francesco (University of California, San 
Diego) ; Marshall Edgell (University of 
North Carolina) ; Sharyn Endow (Yale 
University) ; David Harrison (Jackson 
Laboratory) ; Clyde Hutchinson (Uni- 
versity of North Carolina) ; Lawrence 
Kedes and Lawrence Korn (both of 
Stanford University) ; A. A. Kroon 
(Groningen) ; Richard Novick (Public 
Health Research Institute) ; Tokindo 
Okado (Kyoto University) ; Arnold Rev- 
zin (University of Oregon) ; Robert 
Roeder (Washington University) ; T. 
Thomas (University of Georgia) and H. 
Wahn (Wayne State University). 




Adler, W. H., see Ozato, K. 

Brown, D. D., and R. Stern, Methods of gene 
isolation. Ann. Rev. Biochem., Ifi, 667-693, 

Brown, D. D., see also Brown, R. D., Brownlee, 
G. C, Forsheit, A. B., Lizardi, P., and 
Wellauer, P. K. 

Brown, R. D., and D. D. Brown, The nucleo- 
tide sequence adjoining the 3' end of the 
genes coding for 5S ribosomal RNA in 
Xenopus. Fed. Proc, 34, 608 (abstract), 

Brownlee, G. C., E. M. Cartwright, and D. D. 
Brown, Sequence studies of the 5S DNA of 
Xenopus laevis. J. Mol. Biol, 89, 703-718, 

Carroll, D., see Wellauer, P. K. 

Cartwright, E. M., see Brownlee, G. C. 

Coon, H. G., see Dawid, I. B. 

Davidson, N., see Forsheit, A. B. 

Dawid, I. B., I. Horak, and H. G. Coon, The 
use of hybrid somatic cells as an approach 
to mitochondrial genetics in animals. Genet- 
ics, 78, 459-471, 1974. 

Dawid, I. B., see also Leister, D. E., Wellauer, 
P. K, and Wu, G.-J. 

Deutch, A., see Wellauer, P. K. 

Devreotes, P. N., and D. M. Fambrough, 
Acetylcholine receptor turnover in mem- 
branes of developing muscle fibers. J. Cell 
Biol, 65, 335-358, 1975. 

Devreotes, P. N., see also Fambrough, D. M. 

Ebert, J. D., Birth defects: the present and 
the future. In Birth Defects, A. G. Motulsky, 
W. Lenz and F. J. E. Ebeling, eds. Excerpta 
Medica, pp. 10-17, Amsterdam, 1974. 

Ebert, J. D., see also Ozato, K., and Stam- 
brook, P. J. 

Fambrough, D. M., Acetylcholine receptors: 
revised estimates of extrajunctional receptor 
density in denervated rat diaphragm. /. Gen. 
Physiol, 64, 468-472, 1974. 

Fambrough, D. M., and P. N. Devreotes, Syn- 
thesis and degradation of acetylcholine re- 
ceptors in cultured chick skeletal muscle. In 
Exploratory Concepts in Muscular Dystro- 
phy II, A. T. Milhorat, ed. Excerpta Medica, 
Amsterdam, pp. 55-67, 1975. 

Fambrough, D. M., see also Devreotes, P. N. 

Forsheit, A. B., N. Davidson, and D. D. Brown, 
An electron microscope heteroduplex study 

of the ribosomal DNAs of Xenopus laevis 
and Xenopus mulleri. J. Mol Biol, 90, 301- 
314, 1974. 

Gardner, E., and R. O'Rahilly, Le cerveau a 
la fin de la periode embryonnaire vraie. 
Bull Ass. Anat., 58, 109-114, 1974. 

Gardner, E., see also O'Rahilly, R. 

Gasser, R. F., Atlas of Human Embryos, Harper 
and Row, Hagerstown, Md., 318 pp., 1975. 

Higashinakagawa, T., and R. H. Reeder, Puri- 
fication and preliminary characterization of 
active ribosomal gene chromatin. Federation 
Proceedings, 34, 581 (abstract), 1975. 

Higashinakagawa, T., see also Wellauer, P. K. 

Horak, I., see Dawid, I. B. 

Huang, L., see Pagano, R. E. 

Kelley, D. E., see Wellauer, P. K. 

Leister, D. E., and I. B. Dawid, Physical prop- 
erties and protein constituents of cytoplasmic 
and mitochondrial ribosomes of Xenopus 
laevis. J. Biol. Chem., 249, 5108-5118, 1974. 

Lizardi, P. M., and D. D. Brown, The length 
of the fibroin gene in the Bombyx mori 
genome. Cell, 4, 207-215, 1975. 

Lizardi, P. M., R. Williamson, and D. D. 
Brown, The size of fibroin messenger RNA 
and its polyadenylic acid content. Cell, 4, 
199-205, 1975. 

O'Rahilly, R., and E. Gardner, Les stades de 
developpement de l'embryon humain. Bull. 
Ass. Anat., 58, 177-182, 1974. 

O'Rahilly, R., see also Gardner, E. 

Ozato, K., W. H. Adler, and J. D. Ebert, The 
effect of concanavalinA and bacterial lipo- 
polysaccharides on mixed lymphocyte cul- 
tures and generation of cytotoxic effector 
lymphocytes. Federation Proceedings, 34, 
1005 (abstract), 1975. 

Pagano, R. E., L. Huang, and C. Wey, Inter- 
action of lipid vesicles with cultured mam- 
malian cells. Nature, 252, 166-167, 1974. 

Perry, R., see Wellauer, P. K. 

Ramsey, E. M., The Placenta of Laboratory 
Animals and Man, Holt, Rinehart and Win- 
ston, Inc., New York, 196 pp., 1975. 

Reeder, R. H., Ribosomes from eukaryotes: 
genetics. In Ribosomes, P. Lengyel, M. 
Nomura and A. Tissieres, eds., Cold Spring 
Harbor, 489-518, 1974. 



Reeder, R. H., see also Higashinakagawa, T., 
and Wellauer, P. 

Sachs, H. G., see Stambrook, P. J. 

Stambrook, P. J., H. G. Sachs, and J. D. Ebert, 
The effect of potassium on the cell membrane 
potential and the passage of synchronized 
cells through the cell cycle. J. Cell Physiol., 
85, 283-292, 1975. 

Stambrook, P. J., and R. Williamson, Error 
frequency in 5S RNA from cells grown in 
5-Bromodeoxyuridine. Eur. J. Biochem., 48, 
297-302, 1974. 

Suzuki, E., see Suzuki, Y. 

Suzuki, Y., and E. Suzuki, Quantitative mea- 
surements of fibroin messenger RNA syn- 
thesis in the posterior silk gland of normal 
and mutant Bombyx mori. J. Mol. Biol., 88, 
393-407, 1974. 

Wellauer, P. K., and I. B. Dawid, Secondary 
structure maps of ribsosomal RNA and DNA 
I. J. Mol. Biol, 89, 379-395, 1974. 

Wellauer, P. K., I. B. Dawid, D. E. Kelley, 
and R. P. Perry, Secondary structure maps 
of ribosomal RNA II. J. Mol. Biol, 89, 
397-407, 1974. 

Wellauer, P. K., and R. H. Reeder, A compari- 
son of the structural organization of ampli- 
fied ribosomal DNA from Xenopus mulleri 
and Xenopus laevis. J. Mol Biol, 93, 151- 
161, 1975. 

Wellauer, P. K., R. H. Reeder, D. Carroll, D. 
D. Brown, A. Deutch, T. Higashinakagawa, 
and I. B. Dawid, Amplified ribosomal DNA 
from Xenopus laevis has heterogeneous 
spacer lengths. Proc. Nat. Acad. Sci., USA, 
71, 2823-2827, 1974. 

Wey, C, see Pagano, R. E. 

Williamson, R., see Lizardi, P., and Stambrook, 
P. J. 

Wu, G.-J., and I. B. Dawid, In vitro transcrip- 
tion of Xenopus mitochondrial deoxyribo- 
nucleic acid by homologous mitochondrial 
ribonucleic acid polymerase. J. Biol. Chem., 
249, 4412-4419, 1974. 


Year Ended June 30, 1975 

(including those whose services began or ended during the year) 

Research Staff 

Donald D. Brown, Biochemistry 
Igor B. Dawid, Biochemistry 
James D. Ebert, Director 
Douglas M. Fambrough, Biochemistry 
Kenneth J. Muller, Neurobiology 
Richard E. Pagano, Biophysics 
Ronald H. Reeder, Biochemistry 
Yoshiaki Suzuki, Biochemistry 

Assistant Investigator 
Peter J. Stambrook 

Research Associates (extramural) 

Bent G. Boving, Detroit, Michigan 
Robert L. DeHaan, Atlanta, Georgia 
Ernest Gardner, Davis, California 
Arthur T. Hertig, Boston, Massachusetts 
Irwin R. Konigsberg, Charlottesville, Virginia 
Ronan O'Rahilly, Davis, California 


Ronald D. Brown, Fellow of Carnegie Insti- 
tution of Washington 

Dana Carroll, Fellow of U.S. Public Health 

Scott Emmons, Fellow of the National Cystic 
Fibrosis Research Foundation 

Toru Higashinakagawa, Joint Fellow of Car- 
negie Institution of Washington and Mit- 
subishi-Kasei Institute 

Leaf Huang, Fellow of Carnegie Institution 
of Washington 

John F. Morrow, Fellow of the National 
Cystic Fibrosis Research Foundation 

Seigo Ohi, Fellow of Carnegie Institution of 

Tokindo Okada, Senior Fellow of Carnegie 
Institution of Washington 

Keiko Ozato, Joint Fellow of Carnegie In- 
stitution of Washington and Marine Bio- 
logical Laboratory 



Aileen K. Ritchie, Fellow of Muscular Dys- 
trophy Associations of America, Inc. 

Masatoshi Takeichi, Fellow of Carnegie In- 
stitution of Washington 

Katherine Tepperman, Fellow of the U.S. 
Public Health Service 

William B. Upholt, Fellow of Carnegie In- 
stitution of Washington 

Peter K. Wellauer, Fellow of the National 
Cystic Fibrosis Research Foundation 


Sandra L. Biroc, Graduate, Johns Hopkins 

Peter Devreotes, Graduate, Johns Hopkins 

Marc Friedman, Graduate, Johns Hopkins 

Carol Kaushagen, Graduate, Johns Hopkins 

Jose Ramirez, Graduate, Johns Hopkins 

Changlin A. Wey, Graduate, Johns Hopkins 


Visiting Investigators and 
Extramural Collaborators 

W. H. Adler, Baltimore, Maryland 
George Brownlee, Cambridge, England 
John W. Chase, Boston, Massachusetts 
Hayden G. Coon, Bethesda, Maryland 
N. Davidson, Pasadena, California 
Daniel B. Drachman, Baltimore, Maryland 
Michael Edidin, Baltimore, Maryland 
A. Forsheit, Pasadena, California 
J. W. S. Harris, London, England 
Uel J. McMahan, Boston, Massachusetts 

Christian Thomas, Brussells, Belgium 
Witold Wozniak, Poznan, Poland 

Student Assistant 

Gail Anderson, Johns Hopkins University 

Clerical and Technical Staff 

Elaine S. Asch, Senior Technician 

Judith M. Biemiller, Secretary 

James H. Blackwell, Custodian 

Paul Blackwell, Custodian (part time) 

Julia L. Certain, Librarian (part time) 

Huei-Wen Chen, Technician 

Sally S. Cherry, Technician 

William J. Cleary, Recorder 

William H. Duncan, Senior Technician 

Ernestine V. Flemmings, Laboratory Helper 

Paul Giza, Technician 

Richard D. Grill, Photographer 

Virginia S. Hicks, Laboratory Helper 

Mary E. Hogan, Technician (part time) 

John E. Jones, Custodian 

Eddie D. Jordan, Senior Technician 

June M. Lam, Secretary 

Alice H. Mabin, Laboratory Helper 

Thomas F. Malooly, Business Manager 

Thomas F. Miller, Custodian 

Ann N. Murphy, Secretary 

Joyce I. Patterson, Laboratory Helper 

John Pazdernik, Jr., Building Engineer 

Betty Lou Phebus, Bookkeeper-Clerk 

Martha L. Rebbert, Senior Technician 

Adrienne Robinette, Secretary 

John J. Sexton, Custodian 

Bessie H. Smith, Laboratory Helper 

Delores V. Somerville, Senior Technician 

John L. Wiser, Machinist 

of Terrestrial Magnetism 

Washington, District of Columbia 

George W. Wetherill 

Carnegie Institution of Washington Year Book 74, 1974-1975 


Introduction 109 

Astrophysics 113 

Galaxies 113 

Observational cosmology 113 

Observations of late-type southern galaxies 118 

Internal motions in barred spiral galaxies 120 

The velocity field of the peculiar galaxy NGC 1275 123 

Ionized gas in the nuclear bulge of M31 125 

Large-scale structures in our Galaxy 126 

Search for shell structure in association with old galactic supernova 

remnants in neutral hydrogen and formaldehyde 126 

Zeeman splitting of the 21-cm line of neutral hydrogen 

in supernova remnant shells 129 

Magellanic stream gas 131 

A search for very high velocity clouds 131 

The structure of globular clusters 133 

Equipment development 135 

Image tube systems 135 

The intensifier-image dissector scanner 137 

The pointing of our 60-ft Derwood radio telescope 138 

A new receiver for the IAR 139 

Acknowledgments 139 

Activities of the Instituto Argentino de Radioastronomia 139 

Nuclear and Atomic Physics 141 

Introduction 141 

Continuum and characteristic x rays produced with potassium ions 141 

The 15 N(p,a) 12 C reaction excited with polarized protons 146 

The 9 Be(p,n) 9 B reaction excited with polarized protons 148 

A return to 4 He(p,p) 4 He 149 

Miscellany 150 

Biophysics 150 

Introduction 150 

Properties of human DNA sequences containing poly(dA) or poly(dT) 151 

Relationships of some African monkeys as determined by their single-copy 

DNA sequence homologies 158 

A special class of related repeated sequences in mouse and rat DNAs 159 

Are there mouse satellite sequences in main-band RNA? 161 

Repeated DNA and speciation 164 

A distinctive feature of the human genome 165 

Protein-protein interactions 165 

Historical review of Biophysics Section 172 

Geochemistry and Geophysics 180 

Diffusion of Ca and Sr in a basalt melt 183 

The partitioning of rare earth elements between garnet and liquid 

at high pressures: preliminary experiments 189 

An assessment of local and regional isotopic equilibrium in a 

partially molten mantle 195 

Kinetic processes and thermal history of rocks 210 

Deep Sea Drilling Project Leg 34 basalts and ocean ridge basalts: 

a geochemical comparison 218 

Geochemistry of the Azores and the mid- Atlantic ridge: 29°N to 60°N 224 

Secondary minerals in oceanic basalt 234 

Immiscibility and ancient and modern volcanism 240 

Early evolution of the centra] 'Andean volcanic arc 247 

Strontium isotopic composition of late Cenozoic central Andean volcanic rocks: 

a disequilibrium melting model 250 

Anomalous island arc asthenosphere and continental growth 256 

Determination of the subducting lithosphere boundary by use of converted phases . 266 

Seismic source studies 273 

Archambeau's elastodynamical source-model solution 274 

The implications of moving boundaries in elastodynamics 278 

Multiple rupture earthquakes and the determination of source parameters 281 

Borehole strainmeters : long-term stability and sensitivity to dilatancy 287 

Electrical conductivity studies in South America 291 

Local seismic net in southern Peru 293 

Cosmic-ray research 293 

Bibliography 296 

Personnel 299 


The end of this report year comes at a tributes. This conventional division is 
time of change at the Department of most appropriate to the task of trans- 
Terrestrial Magnetism. The successful mitting existing knowledge from one 
completion of our 30-year program in generation to another. However, in our 
biophysics and the arrival of a new di- case the knowledge is newly generated 
rector represent discontinuities likely to by our work, and we must first prepare 
reopen fundamental questions concerning ourselves to receive this new knowledge 
the long-range goals of the Department, and then assimilate it ourselves before 
It thus seems a good time to reaffirm transmitting this newly won knowledge to 
our commitment to certain approaches to another group of "students," those who 
scientific knowledge which have char- learn from our scientific publications, 
acterized scientific research at DTM. presentations in scientific meetings and 

One of these is a loosely organized from this Report, 
form of group research which occupies 

a middle ground between a disciplined Geophysics and Geochemistry 

"team" with clear lines of authority, and 

a collection of individuals competing The work of the Department in the 

among themselves for individual rewards earth sciences continues to exploit and 

in the form of promotions, financial sup- extend the unification of terrestrial proc- 

port, students, or scientific recognition, esses resulting from the establishment of 

Another is a lack of inhibition in pursu- plate tectonics in the mid-1960's. These 

ing investigations into areas which might processes are most clearly expressed in 

conventionally be regarded as lying out- the oceanic regions of the earth. As a 

side one's field of competence. While the consequence, much of the geochemical 

dangers of this are too obvious to discuss, work in recent years has involved study 

there are also limitations to the extent of oceanic rocks, in contrast to our ear- 

an interdisciplinary team of specialized lier work, which was concentrated almost 

workers can replace a scientist personally entirely on continental material. In 

responsible for all aspects of a scientific 1973-1974 one of our staff members, S. R. 

investigation. We have found that the Hart, served as chief scientist aboard the 

alternative of encouraging mature scien- deep-sea drilling vessel, the Glomar 

tists to enlarge their areas of competence Challenger, on Leg 34 of the Deep Sea 

well beyond their previous bound- Drilling Project, and another of our 

aries can be rewarded by the identifica- group, M. N. Bass, served as a member 

tion of new questions, and by opening up of the scientific staff on that cruise, 
new approaches to older areas of study. According to plate tectonic theory, new 

While this characteristic is not unique to oceanic lithosphere is produced at mid- 

DTM, our endowed financial support is ocean ridges by addition of material from 

especially suited to taking the risks in- deeper within the earth's mantle. The 

volved, avoiding the inhibitions arising surface manifestations of this material 

from a need for short-term results and are the mid-ocean ridge basalts (MORB), 

from the conservatism resulting from in- obtained by dredging and drilling of the 

creasing demands for accountability on sea floor. Previous work, primarily based 

the part of federal funding agencies. on young dredge haul samples, showed 

The Department may therefore be that MORB is distinguishable from other 
viewed as an educational institution of a basalts by both its trace element con- 
special sort, not limited by conventional centrations and its strontium isotopic 
division into faculty and students, even composition. It was not clear from this 
though we do possess these standard at- work whether this was an essential char- 



acteristic of MORB or reflected biased plunge hundreds of kilometers between 

sampling of rocks of a particular age. the continental lithosphere, resulting in 

Hart reports results obtained on older the seismically active Benioff zone, vol- 

(15-40 m.y.) samples of MORB from canism, and the formation of mountain 

the South Atlantic, obtained by drilling, belts. The relationships between these 

which show that the characteristic differ- processes have been the subject of a con- 

ences between MORB and other basalts tinuing combined geophysical and geo- 

extend back into time. chemical investigation. Sacks and Snoke 

These differences between MORB and report seismic studies of lithospheric 

other basalts such as those produced in plate motions in both South America and 

volcanic island arcs are examples of a Japan, while Aldrich reports related 

fundamental problem which dominates studies of electrical conductivity of the 

current geochemical discussions of the earth's interior in the Andes. The seismic 

mechanisms by which material is trans- work suggests a new hypothesis for the 

ferred from the deeper interior of the generation of the thick lithosphere which 

earth to the surface regions. This is that distinguishes continents from ocean ba- 

various basaltic rocks, all presumably sins. This is that the cool, down-going 

derived from the mantle, exhibit different slab of oceanic lithosphere freezes out 

chemical and isotopic characteristics, aesthenosphere in the wedge defined by 

Further examples of this are given in the the intersection of the continental plate 

report by White, Hart, and Schilling con- and the down-going oceanic plate. In 

cerning basalts from the mid-Atlantic combination with migration of the ocean 

ridge, the Azores, and Iceland. The most trench associated with the down-going 

obvious explanation of this is that the slab, this could provide a mechanism for 

mantle is chemically heterogeneous on increasing the mass of continental litho- 

both a local and a global scale. If true, sphere over geological time, 

this opens up a new field of "mantle geol- Investigations reported by James, 

ogy" directed toward an understanding Brooks, and Cuyubamba are directed 

of the basic principles governing these toward extending our knowledge of proc- 

heterogeneities. esses in this region back in time by geo- 

An alternative to the hypothesis of chemical and isotopic investigation of the 

major mantle heterogeneity is that the earliest (185 m.y. old) volcanic rocks of 

differences between the various basalts the Andes. These earliest rocks differ 

are primarily a consequence of chemical from the younger volcanics in that they 

and isotopic disequilibrium on a miner- appear to have been more directly de- 

alogical scale during the production of rived from oceanic lithosphere. The iso- 

the basaltic magma by partial melting of topic compositions of strontium from the 

the mantle. Hart and Hofmann argue younger rocks suggest derivation by 

against this disequilibrium hypothesis as partial remelting of continental litho- 

an explanation of the differences between sphere. More detailed examination of the 

MORB and other basalts. Their reason- strontium isotopic data for the younger 

ing invokes in an essential way the re- rocks shows that it cannot be explained 

suits of new and refined diffusion studies in terms of equilibrium melting of a 

by Hofmann, as well as trace element homogeneous mantle. James presents evi- 

partition studies, such as those reported dence to support a disequilibrium melting 

here by Shimizu and Kushiro. model for these Andean rocks, in spite of 

The tectonic environment of the Andes the arguments of Hart and Hofmann 
and the adjacent portions of the South against a similar model for the genera- 
Atlantic Ocean are complementary to tion of the oceanic basalts, 
those of the mid-ocean ridges. Here A recurring theme in DTM geochemi- 
down-going slabs of oceanic lithosphere cal studies has been the search for in- 


formation concerning the earlier history of galaxies as indicated by both optical 

of the earth, and much of our previous and radio data. This is that the best 

work in geochronology was motivated interpretation of their observations re- 

by that interest. At present the primary quires our Galaxy, as well as its local 

question relating to the older rocks of the neighbors, to be moving at a velocity of 

earth is the extent to which our newly about 500 km/sec with respect to the 

won knowledge of plate tectonics can be reference frame defined by the distant 

extrapolated to much earlier periods of galaxies. The direction of this motion is 

earth history. Earlier work by Hart and perpendicular to the direction of the 

co-workers at DTM supported the idea nearest large cluster of galaxies, the 

that 2.7 b.y. Archean volcanic rocks were Virgo cluster, and the velocity corre- 

in many ways analogous to certain mod- sponds to the orbital velocity of a body 

ern arc volcanics, but that the higher at the distance of our Galaxy from the 

geothermal gradient at that time resulted Virgo cluster. Therefore, the data are 

in a greater proportion of rocks produced consistent with the idea that in some 

by extensive but shallow partial melting, sense our Galaxy is in orbit about the 

In this Report Brooks infers that this Virgo cluster, although not in the usual 

would also produce a larger number of way, as the 10 11 year period of such an 

rocks derived from magmas exhibiting orbit is much greater than the age of the 

liquid immiscibility and presents evi- galaxies. These observations are incon- 

dence that this was the case. sistent with the generally accepted iso- 

The interpretation of seismic data de- tropic distribution of 3°K cosmic black 

pends in an essential way on interaction body radiation background. The dis- 

between observation, experiment, and crepancy appears to be outside the ob- 

physical theory. We have built up our servational errors of both sets of meas- 

strength in the theoretical area in recent urements and represents an important 

years. In this Report, Snoke, Linde, and problem for further investigation. 

Sacks address the problem of relating Many of the optical observations of 

the observed seismograms to the source Rubin, Ford, and Peterson and the radio 

mechanism. They point out major diffi- observations of Thonnard and Turner 

culties with previous solutions to this are directed toward understanding the 

problem and show that it may prove velocity field within a galaxy, particu- 

possible to remove some of these difficul- larly in galaxies of selected types, such 

ties by considering models in which the as the barred spirals and the unusual 

seismic strain release occurs through galaxy NGC 1275, possibly two galaxies 

multiple ruptures rather than a single in collision. Some of this work was done 

rupture. at the radio astronomical facility of the 

Instituto Argentino Radioastronomia 

Astrophysics (IAR) in Argentina, established jointly 

In our work in astrophysics, combining b ^ the Carnegie Institution, Argentine 

both radio and optical observations, at- universities, and Argentine scientific 

tention is centered on distant galaxies agencies. In addition, Kumar has studied 

and their motions, on dynamical prob- the °P tlcal emission spectrum of a 

lems in our own Galaxy, and in the de- "bulge" of ionized gas extending outward 

velopment of the advanced instrumenta- from the nucleus of the nearby Andro- 

tion which will be needed for future meda galaxy, M31. These measurements 

extension of this work. support the hypothesis that this gas is 

Rubin, Ford, and Thonnard have ionized by extremely hot stars near the 

drawn a very interesting conclusion from end of stellar evolution which are in the 

their study of the apparent anisotropy process of becoming white dwarfs, 

in the distribution of a particular class Within our own Galaxy, radio studies 



have been made of supernova remnants. 
Supernovae and their remnants are of 
great interest because of their relation- 
ship to the final stages of stellar evolu- 
tion, to pulsars and neutron stars, their 
major contribution to the synthesis of 
the elements, as a mechanism for return- 
ing stellar matter to the interstellar 
medium, and as sources of cosmic rays. 
Assousa, in collaboration with others, has 
searched for the radio frequency lines 
of formaldehyde in expanding neutral 
hydrogen shells associated with these 
remnants. Although formaldehyde ab- 
sorption was detected, their work showed 
that this was neither associated with the 
expanding shells nor with the remnants 
themselves. Turner has made use of Zee- 
man splitting in the 21 -cm hydrogen line 
of the expanding shells to measure the 
strength of the interstellar magnetic field. 
In one remnant he showed that this field 
was less than 3 microgauss, whereas in 
another case a possible positive result of 
about 5 microgauss was found. 

In other work directed toward our 
own Galaxy and its neighborhood, Tuve 
and his co-workers have searched for 
very high velocity hydrogen clouds which 
would be indicative of the Galaxy being 
an open system, either receiving material 
from intergalactic space or expelling 
material from the Galaxy. No such phe- 
nomena were found. Peterson reports 
work on a very distant globular cluster 
associated with our Galaxy. Comparison 
of the size of the cluster with the tidal 
"Roche limit" for disruption by the gen- 
eral galactic gravitational field is used 
to show that this cluster is in a very 
eccentric orbit about the galactic center. 
Many of the observations mentioned 
above would not have been possible with- 
out the Carnegie image tube developed 
at DTM by W. K. Ford, Jr. Continued 
improvements in these devices and sys- 
tems for utilizing them have been made 
by Ford. In addition he reports advances 
in developing a new type of detector 
especially suitable for photometric work. 
This detector eliminates use of the pho- 

tographic plate and provides the imaging 
data in digital form. 


Although it is planned that some work 
in biophysics will continue beyond this 
year, it will be on a diminishing scale. It 
therefore seemed a good time to sum- 
marize the many distinguished achieve- 
ments of this group during the past 30 
years. This has been done by R. B. 
Roberts, one of the founding members of 
the group. He traces the evolution of the 
DTM Biophysics Section from a group 
of enthusiastic nuclear physicists with 
little background in biology through 
their increasing biological sophistication, 
and enumerates 18 significant findings 
made by the nine staff members and 
numerous resident associates, fellows, 
and technical assistants associated with 
this program. All this work was done in 
the context of the revolutionary advances 
being made in biochemistry during this 
time, as a consequence of new under- 
standing of the genetic mechanism in 
living cells and the role of nucleic acids. 
It is a testimony to the viability of the 
approach to science taken by the Insti- 
tution and the Department that a small 
group with modest funding and few pro- 
fessional credentials in this area were 
able to play such a significant role in a 
large and expanding field of science. 

During this year studies of DNA in 
human, primate, rat, and mouse cells 
have continued. Bonner has worked on 
a promising method for isolating geneti- 
cally important human DNA segments. 
Hoyer has used the thermal stability of 
reassociated DNA obtained from differ- 
ent primates to determine the relative 
times of their evolutionary divergence. 
By use of this method he infers that the 
group of African monkeys studied repre- 
sents an evolutionary group more homo- 
geneous than that formed by man, orang- 
utan, and gibbon but less homogeneous 
than man, chimpanzee, and gorilla. 

N. Rice and other members of the 
group have been working on the relation- 


ship between repeated DNA sequences on the atomic number of the target atom, 

(families) and evolutionary processes. Surprisingly, it appeared that current 

In Rice's work the puzzling result is theories do not yield this result, even 

found that two species can possess homo- though its simple form strongly suggests 

geneous families which are only loosely that a theoretical explanation should be 

related to each other, and alternative possible. Brown and his co-workers have 

hypotheses are proposed to explain this derived a simple expression which fits 

result. Roberts and co-workers report the observed exponential dependence of 

preliminary work supporting the hy- the line spectra on Z, as well as predict- 

pothesis that family formation is neces- ing the x-ray energy spectrum of the 

sary for speciation. continuum radiation. Although this does 

Cowie is working at present at the not yet represent a complete theory for 

Pasteur Institute in Paris, continuing the problem, the agreement with ob- 

work on metabolism in Escherichia coli. served data indicates that the approach 

He reports results on the modification of taken is on the right track, 

enzyme activity as a consequence of In the area of nuclear physics, the 

enzyme-antibody binding. nuclear structure of 16 and 10 B was 

. successfully investigated by formation of 

Nuclear and Atomic Physics the compound nucleus by bombardment 

One of the older fields of research still of 15 N and 9 Be with polarized protons, 

very active at DTM is the study of The former nucleus is of astrophysical 

atomic and nuclear phenomena by bom- interest as 15 N(p,«) 12 C and 15 N(p,y) 16 

bardment with accelerated charged ions, are competing steps in the CNO cycle 

This work is carried out by L. Brown of stellar energy production, 

and his colleagues G. H. Pepper and The results found for the energy levels 

L. G. Arnold, and also involves collabo- of 10 B by 9 Be(p,n) 9 B are satisfying in 

ration with several workers associated that they are in agreement with the pre- 

with our astrophysics group, N. Thon- diction of the shell model as well as with 

nard and C. K. Kumar. some earlier work. However, they are in 

In the atomic physics area, the inten- disagreement with our earlier measure- 

sity of production of characteristic x-ray ments of 9 Be(p,p) 9 Be. The latter data 

line spectra by heavy ion bombardment are being reanalyzed in order to try to 

yielded a simple exponential dependence understand this difference. 


G. E. Assousa, E. T. Ecklund, W. K. Ford, Jr., C. K. Kumar, C. A. Little, C. J. Peterson, 
V. C Rubin, N. Thonnard, K. C. Turner, M. A. Tuve, and J. W. Warner 

B. Balick, L. K. De Noyer, J. W. Erkes, J. Graham, C. Heiles, M. S. Roberts, and R. Sanders 

Astronomers at DTM are principally GALAXIES 

observers, using optical or radio tele- Observational Cosmology 
scopes as the problems demand, and 

turning to the development of more ad- w - Kent Ford > Jr - Vera c - Rubin, 
vanced instruments as the need for them and Norbert Thonnard 
is indicated. Major studies of the past For the past several years, we have 
year relate to external galaxies and large been collecting data for an all-sky sam- 
structures in our Galaxy, and are dis- pie of Sc I galaxies, i.e., ''standard can- 
cussed below. dies," to answer the question, Is the 



Hubble expansion isotropic as observed 
from the Galaxy? Curiously, the answer 
to this question had not explicitly been 
sought previously in an observational 
program by optical astronomers. In col- 
laboration with Dr. Morton S. Roberts, 
National Radio Astronomy Observa- 
tory,* and Dr. John A. Graham, Cerro 
Tololo Inter-American Observatory, we 
have now obtained optical velocities, 21- 
cm velocities, HI fluxes, magnitudes, and 
colors for almost 200 galaxies, and the 
analysis of the observed expansion is 
presently being completed. The anisot- 
ropy reported earlier (Rubin, Ford, and 
Rubin, 1973; Year Book 73, 885, 1974) 
persists at about the 4.5 o- level. From a 
detailed study of the form of the velocity 
variation across the sky for galaxies with 
3500 < V < 6500 km s- 1 (distances 
about 60 to 120 Mpc), we believe that 
the apparent isotropy can best be in- 
terpreted as a motion of our Galaxy with 
respect to this sample of galaxies. The 
velocity of the sun is V = 647 ± 144 km 
s- 1 toward a = 39° ± 24°, 8 = +57° 
±12°, which corresponds to galactic co- 
ordinates I = 137°, b = —2°. If we 
remove from this motion the velocity of 
our sun relative to the Local Group of 
galaxies, then the velocity of our Galaxy 
and the Local Group of galaxies becomes 
V = 494 ± 150 km s" 1 , toward I = 
164°, b = —3°. 

This location for the apex of the galac- 
tic motion means that the Galaxy is 
moving edge on, and that the motion is 
not far from the plane defined by the 
distribution of bright galaxies and small 
nearby groups, the Supergalactic Plane. 
Moreover, the motion is approximately 
orthogonal to the direction of the Virgo 
cluster, the nearest large cluster of 
galaxies, of which we may be an outlying 
member. Hence we have no component 
of motion toward or away from the Virgo 
cluster, nor has the presence of the excess 

* The National Radio Astronomy Observa- 
tory is operated by Associated Universities for 
Research in Astronomy, Inc., under contract 
with the National Science Foundation. 

mass toward Virgo slowed the expansion 
in our region of space. A schematic dia- 
gram of this motion is shown in Fig. 1. 
Our program galaxies are too distant to 
enable us to tell if our Local Group is 
rotating about the Virgo cluster. At our 
present distance from Virgo, about 20 
Mpc, a bound orbit would have V ~ 500 
km s -1 , with a period of revolution of 
some 10 11 years, or ten times the esti- 
mated age of the universe. Hence the 
concept of a bound orbit has no meaning, 
unless it relates to an earlier age of the 
expanding universe, when our Galaxy 
was closer to Virgo. From our study we 
cannot define the size of our co-moving 
group, nor draw any conclusions con- 
cerning systematic motions of the gal- 
axies nearer than 60 Mpc. We initially 
chose a sample for study which is suffi- 
ciently distant so that the velocities are 
unaffected by local irregularities within 
the local supercluster, yet not so distant 
that any velocity irregularity is masked 
by the large Hubble expansion. 

A motion of our Galaxy as large as 
647 ± 144 km s _1 as observed from the 
sun is of considerable interest to cosmol- 
ogists because it conflicts with present 
determinations of the isotropy of the 3° 
black body radiation. The early history 
of the universe is contained in the spec- 
trum and the directionality of the 3° 
cosmic background radiation. The most 
generally accepted theory assumes that 
the universe started some 15 billion years 
ago at a singularity, the Big Bang, and 
has been expanding and cooling ever 
since. At present, the radiation field from 
the Big Bang has cooled to 3°K; the 
corresponding peak wavelength of the 
photons is about 1 or 2 mm. To study 
the isotropy of this radiation field from 
the earth is very difficult because of the 
radiation of the earth's atmosphere and 
the Galaxy, which are both considerably 
above 3°. It is only slightly less difficult 
from balloons. However, 'results by a 
number of groups (discussed in Peebles, 
1971) indicate that the radiation field is 
extremely isotropic, from which an upper 




Fig. 1. Schematic drawing in plane of supergalaxy. Our Galaxy is shown moving edge on 
toward SGL = 0°, approximately orthogonal to the Virgo cluster. The Sc I galaxies in the 
study are located in a spherical shell with 3500 < V < 6500 km s" 1 . 

limit to the motion of the sun of about 
300 km s _1 is inferred. If future experi- 
ments confirm the isotropy of the radi- 
ation field, then the resolution of the 
conflict between the velocity we have 
detected and the low velocity inferred 
from the 3° radiation may be a major 
step in our understanding of the universe. 
Alternative explanations of the ob- 
served anisotropy might be a Hubble 
constant which varies as a function of 
position on the sky, or a mean absolute 
magnitude for Sc I galaxies which varies 
as a function of position. In either case, 
however, the observations require a mini- 
mum value on one region of the sky, 
increasing smoothly to a maximum value 
180° away. Such a cosine variation is 
the natural result of a motion, and hence 
we feel this latter explanation is the most 
acceptable. Velocity studies of a more 
distant sample of galaxies could distin- 

guish between these two interpretations, 
and we are making plans to extend the 
study to both nearer and more distant 

Enroute to the conclusion of our Sc I 
study, we used the extensive body of 
data to carry out a considerable number 
of statistical analyses, much in the spirit 
of early studies of statistics of stars. 
From the variation of velocity, magni- 
tudes, and colors of the galaxies, we 
mapped the large-scale distribution of 
absorbing material in our Galaxy. This 
is the first time that a complete sample 
of galactic velocities has been employed 
in such a determination, and the result is 
a completely independent determination 
of the coefficient of galactic extinction, 
A B . We obtain A B — 0.14 ± 0.04, which 
supports the low value for galactic ex- 
tinction which is derived from stellar 



dnoj6 |doo| 







. "I 

•*• • • 

t • • 

• • • 

• • • • • 

•• • • • 

• • .» • \ • • • y . 
• • 1 • 

• • • 

• •• 
• • • • •. 


• • • 

• • • 

• •• • 

•*•••!• t 

•• • •» 

• • • • • • 

• • • • • 

• M • • 

• •• • • • • - • 

• • t* 

• •• • •• • 

• • t • • 

• • • • 

• • • 

• • • • • 

• • 







, dno>i6 idooi , , 

S UU^) ' ' A|IOO|9A DO "| 






-o o 

~ *_ 

c o 





CD <D 



o ^ 





c o 

§ o 
£- o 

O M- 









p 5 

bfi <l> 

5 S 

-a O 


•■H *0 
CO .+J 


M bfi 



° a 

d « 

.2 -B 

-l-=> 03 

o -d 







u d 

CD -d 
£ OQ 


o . 
o ^ 

"vU o 

> P 

> -^ 

D -1-3 

tn c3 



bfi a 

1-1 a) 

w o 

2 T3 

J o 
bfl £ 

a§ d 

eS ~ O 

pv| g 

o SP& 

* 3"S 

tB d O 

S S S 

bC u d 
^ CD ?T 

id M t^ 

^^ s 

• <^ 5s 
«.« | 

r? *» 

Pi -t-3 



We have also detected a very strict studies that all Sc I galaxies have the 

relation between intrinsic luminosity of same mean absolute magnitude, M = 

an Sc I galaxy and its diameter, a rela- M , with a dispersion of g Mq ~ 075. Now, 

tion which can be used to assign a more however, by considering the linear diam- 

exact luminosity to each galaxy. For- eter of the galaxy, we can make a better 

merly, it had been assumed in statistical estimate of the absolute magnitude of 


D 83 

i r 

s^~ J \< 

D 87 



"i i r 

D 89 

' V -" N -V V/ 




1 1 



D 92 



tl 1 


D 93 


-< =■ — v ■ 



1 1 1 

. r\ , 

D 95 



D 96 



I I 


D 86 


i ti 



D 97 



D 98 




1 I 

D 100 





~i i r 

D 102 


^\f^J^-^ - 



fjVW^^ v/"A/v 

D 103 

v^ - A/WW, 

2.5 MHz 



D 106 




cAX ,. -l. 

-r — (km s ) — "- 


Fig. 3. Representative sample of 21-cm hydrogen line velocity profiles for Sc I galaxies. The 
abscissa gives the 21-cm flux in Janskies (1 Jy = 10" 26 w m~ 2 Hz -1 ), while the ordinate is the 
heliocentric velocity in km sec" 1 . The arrows indicate the observed optical velocities. 



each galaxy. This enables us to get a 
more accurate estimate of the distance 
of each galaxy by comparing its ob- 
served magnitude with its assigned in- 
trinsic luminosity. The variation is such 
that an Sc I galaxy with diameter 20 
kpc is intrinsically 0.7 magnitudes fainter 
than one of 40 kpc ; one of 60 kpc is m 4 
brighter. While the relation between gal- 
axy diameter and luminosity has been 
discussed several times since it was 
explicitly noted by Hubble (1936), no 
use of the relation had been made. We 
show in Fig. 2A and 2B the Hubble 
diagram for our program galaxies, before 
and after application of the correction 
for the diameter/luminosity effect. The 
very significant decrease in the scatter 
(<r M < m 4) of the diagram is due to the 
application of this correction. 

We are analyzing our 21 -cm neutral 
hydrogen line observations of the Sc I 
sample and therefore can only mention 
some preliminary results. The success of 
the observations in itself is impressive. 
Of 137 galaxies observed, in only two 
were we not able to detect any hydrogen 
emission. In five galaxies the signal was 
confused by a companion galaxy in the 
telescope beam, and for five others the 
detection was questionable and was not 
included in our analysis. We therefore 
have 125 Sc I galaxies with a good de- 
termination of the total hydrogen flux. 
Because these galaxies are all much 
smaller in diameter than the telescope 
beam, no errors should arise due to partly 
resolved sources. The analysis of the 
hydrogen contents of such a large, well- 
defined set of galaxies should be invalu- 
ble in helping us to better understand the 
properties of spiral galaxies. 

In Fig. 3 a representative sample of 
hydrogen line profiles is shown for 16 Sc 
I galaxies. The arrow indicates the ob- 
served optical velocity. For the 120 gal- 
axies with both optical and 21 -cm veloci- 
ties, the mean difference F opt — V 21 om is 
only 3 km/sec, with o- = 36 km/sec. The 
ratio of optical luminosity to neutral 
hydrogen luminosity varies by no more 

than a factor of 3, indicating that the 
neutral hydrogen luminosity can also be 
used as an indicator of the "brightness" 
of the galaxy. We therefore use the 21-cm 
luminosity as a second distance indi- 
cator, in the same manner as the optical 
magnitudes. Because galactic extinction 
will not affect the 21-cm magnitudes, the 
close correlation of the optical and 21-cm 
luminosities confirms the validity of the 
correction for absorption which we use, 
and, more importantly, helps to rule out 
galactic absorption as a source of the 


Hubble, E., The Realm of the Nebulae, 
New Haven: Yale University Press, 
p. 48, 1936. 

Peebles, P. J. E., Physical Cosmology, 
Princeton: Princeton University Press, 
Ch. 2, 1971. 

Rubin, V. C, W. K. Ford, Jr., and J. S. 
Rubin, A curious distribution of radial 
velocities of Sc I galaxies with 14.0 ^ 
m ^ 15.0, Astrophys. J. (Lett.), 183, 
Llll, 1973. 

Observations of Late-type 
Southern Galaxies 

Norbert Thonnard and Kenneth C. Turner 

Of importance in understanding the 
contents and structure of galaxies has 
been a comparison of their neutral hy- 
drogen content and velocity distribution 
with their optical morphology. From the 
width of the hydrogen line profile and 
the inclination of the galaxy on the plane 
of the sky, it is possible to obtain an 
independent determination of the total 
mass of the galaxy. The ratio of hydro- 
gen to optical luminosity and the total 
hydrogen mass are important parameters 
in understanding the properties of differ- 
ent galaxy types. The systematic study 
of large numbers of galaxies at 21 cm 
has until recently been undertaken only 
from northern hemisphere observations. 
In contrast, the Australian observatories 
have focused on the detailed study of the 



velocity field of individual galaxies that 
were large enough to be resolved by their 
13V2 arc min beam (Shabbrook and 
Robinson, 1967). Observations of gal- 
axies at 21 cm from the southern hemi- 
sphere, especially with telescope beams 
that are large compared to the galaxy 
size, are important to further our under- 
standing of galaxies. 

We therefore continued our program 
of 21-cm hydrogen line observations of 
late-type galaxies in the southern sky 
following the installation of a redesigned 
receiver front-end on the 30-meter CIW- 
IAR radio telescope in Parque Pereyra, 
Argentina. Even though the receiver 
noise temperature had improved by 25%, 
we still had difficulty observing these 
faint signals because the system stability 

and base-line curvature continued to be 
a problem. To remove the instrumental 
base line (approx. 8°K), we had to spend 
6 minutes out of every 12-minute inte- 
gration cycle observing a point away 
from the galaxy which we later sub- 
tracted from the "on" profile. 

We observed 11 galaxies. Two of these, 
in which we had detected 21-cm hydro- 
gen line emission for the first time during 
our first observing run {Year Book 71, 
p. 243, 1972), were large enough that we 
mapped them this time to determine the 
total extent of the neutral hydrogen. In 
Fig. 4 are shown representative profiles 
of three galaxies. Typical integration 
times for each object averaged about 90 
minutes "on" and 90 minutes "off." 

A-255 is a late barred spiral, classified 






A 255 

<x=2 h 55.l m 
8 = -54° 48' 

8 =-41° 17' 

Optical velocity 

NGC 2997 
o<= 9 h 43.5' 
8 =-30° 58' 



Optical velocity 
J I L 

175 300 425 550 675 800 925 1050 

Radial velocity (km/sec) 


Fig. 4. 21-cm neutral hydrogen line profiles for three southern galaxies. The abscissa gives 
the antenna temperature in degrees Kelvin; the ordinate gives the observed velocity corrected 
to the local standard of rest in km/sec. Where known, published optical velocities are 



as an SBd by de Vaucouleurs and de 
Vaucouleurs (1964). This galaxy has no 
published optical or 21-cm velocity; 
therefore, the radial velocity of 555 km/ 
sec is a new value. 

NGC 1291 is an interesting galaxy. 
Perrine (1922) described it as a "fuzzy 
ball" with a faint outer spiral structure 
of approximately one and one half turns. 
In the Shapley-Ames catalog it is classi- 
fied as an elliptical and by de Vaucouleurs 
as a ringed barred SO/a galaxy. De 
Vaucouleurs (1963) obtained an optical 
radial velocity of 802 km/sec. On the 
Palomar Sky Survey prints, except for 
the faint outer ring, this object looks 
like a normal barred SO galaxy. It has 
been found that as one goes along the 
Hubble sequence from irregular (late- 
type) to elliptical (early-type) galaxies, 
the fraction of matter in the form of 
neutral hydrogen as compared to total 
galaxy mass decreases to vanishing 
amounts; no 21-cm hydrogen line radi- 
ation has yet been found in elliptical 
galaxies, and the detection for "normal" 
SO's is still debatable (Gallagher, Faber, 
and Balick, 1975) . All of this would make 
it quite surprising to find neutral hydro- 
gen in this galaxy, but the good agree- 
ment between the original detection of 
21-cm emission with V r = 833 km/sec 
(Lewis, 1970), our present observation 
of V r = 828 km/sec and the optical 
velocity of V r = 802 km/sec makes the 
detection of hydrogen emission from 
NGC 1291 quite certain. Whether the 
surprisingly high neutral hydrogen con- 
tent warrants reclassification of this gal- 
axy or is a characteristic of the ringed 
structure, can only be settled by more 
observations on similar objects. 

NGC 2997 is a typical Sc galaxy simi- 
lar to M101. An optical spectrum taken 
by Pastoriza (1967) shows a rather steep 
velocity gradient, not at all inconsistent 
with our 21-cm profile. Also, using an 
inclination of 24° determined from the 
ratio of major to minor axes and a typi- 
cal maximum rotational velocity of 275 
km/sec for an Sc galaxy, one calcu- 

lates a profile width of 224 km/sec, 
which is in good agreement with the 
observed width of approximately 250 


de Vaucouleurs, G., and A. de Vaucou- 
leurs, Classification and radial veloci- 
ties of bright southern galaxies, Mem. 
R. Astron. Soc, 68, 69, 1963. 

de Vaucouleurs, G., and A. de Vaucou- 
leurs, Reference Catalogue of Bright 
Galaxies, Austin: University of Texas 
Press, 1964. 

Gallagher, J. S., S. M. Faber, and B. 
Balick, H I in early-type galaxies. I. 
observations. Submitted for publica- 
tion, Astrophys. J., 1975. 

Lewis, B. M., Neutral hydrogen in the 
SO galaxy NGC 1291, Observatory, 
90, 264, 1970. 

Pastoriza, M. G., Spectra of peculiar 
nuclei in galaxies, Observatory, 87, 
225, 1967. 

Perrine, CD., Notes on four interesting 
nebulae, Mon. Not. R. Astron. Soc, 
82, 486, 1922. 

Shabbrook, R. R., and B. S. Robinson, 
21-cm observations of NGC 300, Aust. 
J. Phys., 20, 131, 1967. 

Internal Motions in Barred 
Spiral Galaxies 

W . Kent Ford, Jr., Charles J. Peterson, 
Vera C. Rubin, and Norbert Thonnard 

In the past 15 years, rapid progress 
has been made in understanding the dy- 
namics of spiral galaxies, principally 
because of advances in observing tech- 
nology. For nearby galaxies, the velocity 
field can be mapped in detail by meas- 
uring the line-of-sight velocities of the 
excited gas surrounding the hot, young 
blue stars. Unfortunately, the observa- 
tion and understanding of barred spiral 
galaxies have remained difficult, princi- 
pally because the bars consist of old, 
relatively faint cool stars with little or 
no excited gas. Hence, the velocity field 



within the bar and its relation to spiral 
structure in general remain unknown. 
With the availability of the 4-meter 
telescope and advanced image tubes, we 
are attempting to learn more about the 
dynamics of barred spiral galaxies. 

NGC 5383 is one of the few bright 
barred spiral galaxies in the northern 
sky, and therefore it is a natural candi- 
date for observation in our continuing 
program to study the velocity fields in 

barred spiral galaxies. We have used a 
Carnegie image tube on the Kitt Peak 
4-meter spectrograph to obtain spectra 
at very high spatial and velocity resolu- 
tion in NGC 5383. We show in Fig. 5 a 
photograph of the galaxy taken with a 
Carnegie image tube at the Cassegrain 
focus of the 4-meter telescope, and 
spectra in three position angles across 
the galaxy. The emission lines of Ha, 
[N II], and [S II] arise from the ex- 

NGC 5383 



I I I 


N I 

I I 


I I 



Fig. 5. Reproduction of a plate of NGC 5383, taken at the Cassegrain focus of the Kitt Peak 
4-m telescope with a Carnegie image tube, plus N 2 -baked Illa-J plate, exposure 12 min. Lines 
indicate positions of spectrograph slit for spectra a, b, c. All spectra taken with Carnegie image 
tube plus 4-m Cassegrain spectrograph, N 2 baked Illa-J plates, exposures 125, 85, and 107 min ; 
original dispersion 56 A/mm. Comparison spectral lines are Ne + A. Lines completely crossing 
spectra are emission lines of the night sky. 



cited gas surrounding hot young stars, 
and a study of their velocities allows us 
to map the velocity field of the excited 
gas. We now have observed emission 
line spectra with the spectrograph slit 
in nine different positions. 

In order to determine the velocities of 
the stellar population in the bar, we 
have obtained spectra along the bar in 
the blue spectral region. We are presently 
completing the measurement and reduc- 
tion of this extensive plate material. In 
conjunction with the study, the Dutch 
radio astronomers are using the Wester- 
bork (Holland) aperture synthesis radio 
telescope at 21-cm to map the velocity 
field of the neutral hydrogen in NGC 
5383. Intercomparison of the optical and 
radio results is expected to be valuable 
in helping us to understand the dynamics 
of barred spiral galaxies. 

NGC 3351 is a bright barred spiral 
galaxy in the Leo group of galaxies, typi- 
cal of those barred spirals in which the 
outer arms form a complete ring like the 
Greek letter theta. The galaxy is oriented 

on the sky so that our line of sight is 
almost along the bar. Hence, noncircular 
motions along the bar should be detecta- 
ble. Our study of the nucleus of this 
galaxy is now complete (Rubin et al., 
1975a) and shows the nucleus to consist 
not of a single central concentration but 
a ring of emission regions about 350 pc 
from the center. The ring is rotating, 
with V rot = 126 ± 16 km s _1 , and also 
contracting toward the center, with y con 
= 34 ± 11 km s" 1 . 

With the Kitt Peak 4-meter telescope, 
we have obtained spectra with the spec- 
trograph slit tangent to the outer ring 
in 12 position angles. Figure 6 shows the 
measured hydrogen emission line veloci- 
ties around the ring, plotted as a function 
of angular position in the plane of the 
galaxy. If the outer arms are rotating 
only, the observed velocities will vary as 
a sinusoid, with a phase defined by the 
galaxy /observer geometry and an ampli- 
tude proportional to the rotational veloc- 
ity. The dashed line is a least-squares 
sinusoidal fit which gives a heliocentric 

\ 800 


O 700 - 

T 1 1 1 1 

60" < r< 80" 

— i 1 1 1 1 — 

1 1 — 

i 1 i 1 — 

— i i 1 — 

D 80"<r<100" 

A r > 120" 




□ V 

° o 



\ o 

o ''••.. 



1 1 1 1 — 


dOo Q.--"n 
o .••' 

°.Dfa A fl 

D □ 


o 6\ 


.,"'o° o° 

A A.--" © A 

A<y □ 


_i_ i i i i 

—1 1 1 1 1 

1 ' 

i 1, I 

* o"°,.. A 

° ng 
i i i i 



i I I 



Fig. 6. Observed velocities as a function of angular position in the plane of the galaxy NGC 
3351. Data are divided into four groups according to radial distance from the nucleus. The 
dotted line is a least-squares sinusoidal fit with systemic velocity 775 km/sec and amplitude 
V(R) sini — 139 ± 4 km/sec, corresponding to V = 216 ± 6 km s _1 in the plane of the galaxy, 
3 < R < 6 kpc. Velocity data for 0° < 6 galaxy < 60° has been repeated at the right edge 
of the graph. The two arrows mark the position angle of the bar. 



systemic velocity of 780 ± 3 km s _1 , in 
excellent agreement with the optical and 
21-cm central velocities of 779 ± 8 and 
779 ± 3 km s -1 . The rotational velocity 
is V{R)sin i = 139 ± 4 km s _1 cor- 
responding to a rotation in the plane of 
the galaxy V rot = 216 ± 6 km s _1 at 
the distance of the ring, 3 < r < 6 kpc ; 
the phase indicates that there are no 
significant expansion or contraction mo- 
tions in the outer ring, but rotation only. 
The mass within 9 kpc is ^ 5 X 10 10 M . 

To understand the dynamics of barred 
spiral systems, it is necessary to know 
the velocity structure of the bar. How- 
ever, with one possible exception, there 
are no emission regions in the bar. Meas- 
urement of relatively sharp stellar ab- 
sorption features (Ca I A 4226, Fe I 
4325, Hy, H/?, Mg I A 5183, plus other 
lines) shows that the rotation within the 
bar is of uniform angular velocity. The 
small velocity gradient observed along 
the bar is consistent with only circular 
motion in the bar. The study is con- 

NGC 6764 is a relatively nearby, but 
previously unstudied, barred spiral gal- 
axy. A 4-m spectrum revealed that the 
nuclear emission lines are about 750 km 
s _1 broad, which places it in the class of 
Seyfert galaxies, those galaxies with 
rapidly rotating clouds of ionized gas in 
their nuclei. Measured velocities just 
outside the nucleus indicate that there 
are streaming motions both into and out 
of the nucleus. The details of the velocity 
field have been published elsewhere 
(Rubin et al, 19756). 


Rubin, V. C, W. K. Ford, Jr. and C. J. 
Peterson, Evidence for contraction in 
the nuclear ring of the barred spiral 
galaxy NGC 3351, Astrophys. J., 198, 
39, 1975a. 

Rubin, V. C, N. Thonnard, and W. K. 
Ford, Jr., Observations of NGC 6764, 
a barred Seyfert galaxy, Astrophys. J., 
198, 31, 19756. 

The Velocity Field of the 
Peculiar Galaxy NGC 1275 

W. Kent Ford, Jr., Charles J . Peterson, 
and Vera C '. Rubin 

NGC 1275 is one of the most peculiar 
galaxies known. Ever since the discovery 
that it is an intense source of radio radi- 
ation, astronomers have been attempting 
to understand its nature. Its structure 
has been variously interpreted as the 
result of a collision between two galaxies 
(Baade and Minkowski, 1954) or the 
aftermath of a violent outburst in the 
nucleus of a single galaxy (Burbidge, 
Burbidge, and Sandage, 1963). The ex- 
cited gas appears at two velocities, one 
at V ~ 5200 km s _1 in agreement with 
the velocity determined from the stellar 
absorption lines (Humason, 1932, and 
our spectra), and the other at V ~ 8200 
km s _1 . Interest in this object has been 
restimulated by the recent discovery 
(deYoung, Roberts, and Saslaw, 1973) 
of 21-cm hydrogen absorption, with a 
velocity corresponding to the high veloc- 
ity system. If this is interpreted as two 
galaxies in collision, then the galaxy at 
V ~ 8200 km s _1 must lie along the 
line of sight between us and the galaxy 
at V ~ 5200 km s _1 , and must be falling 
into the lower velocity galaxy. 

We have obtained new observational 
material with the Kitt Peak 4-meter 
image tube spectrograph to map the 
velocity field at higher spatial and veloc- 
ity resolution than has been done for- 
merly to see if an optical study can help 
clarify the nature of NGC 1275. This 
higher resolution reveals much greater 
structure in the velocity field than had 
been previously found (Burbidge and 
Burbidge, 1965). A spectrum taken 7"N 
of the nucleus is shown in Fig. 7; Ha in 
the two velocity systems is marked. The 
very high degree of complexity in the 
relative line strengths of the two sys- 
tems, in the velocity gradients, in the 
discontinuities in the lines, and even in 
the discontinuities in the weak con- 
tinuum, all suggest that it will not be 













5000 ■ 

5200 KM/SEC) {8200 KM/SEC 




tO I] 6300 

D [OI] 6363 

C TNE] 6548 

9 [Nil] 6583 

o Ha 6562 


A [SI] 6717 


V [SE] 6731 




So 9 

o , 


o o 


Sfc 2 8 


-100 -90 -80 

-70 -60 -50 -40 -30 


Fig. 7. Emission line velocities in NGC 1275 measured on 4m-583; spectrograph slit was 
offset ~7"N at position angle 92 ?5. Reproduction of spectrum is shown above, taken with 
Carnegie image tube plus 4-m Cassegrain spectrograph, N 2 -baked Illa-J plates, exposure 230 
min; original dispersion 56 A/mm. The location of Ha in the two velocity systems is marked 
by arrows. 

easy to understand this object. Our ob- Burbidge, E. M., G. R. Burbidge, and 

servations are continuing. 


Baade, W., and R. Minkowski, On 
the identification of radio sources, 
Astrophys. J., 119, 215, 1954. 

A. R. Sandage, Evidence for the oc- 
currence of violent events in the nuclei 
of galaxies, Rev. Mod. Phys., 35, 947, 
Burbidge, E. M., and G. R. Burbidge, 
Optical evidence suggesting the oc- 



currence of a violent outburst in NGC 
1275, Astrophys. J., llfl, 1351, 1965. 

de Young, D. S., M. S. Roberts, and W. 
C. Saslaw, Detection of 21 -centimeter 
hydrogen absorption in the high veloc- 
ity component of the radio galaxy 
Perseus A, Astrophys. J., 185, 809, 

Humason, M. L., The emission spectrum 
of the extra-galactic nebula NGC 
1275, Publ. Astron. Soc. Pac, U, 267, 

Ionized Gas in the Nuclear Bulge 

C. Krishna Kumar 

The relation of the chemical composi- 
tion of the interstellar material in a 
galaxy to the past history of the evolu- 
tion and dynamics of stars in the galaxy 
continues to be a major question in the 
understanding of the universe. Only for 
nearby galaxies is there sufficient spatial 

resolution to permit the investigation of 
these questions as a function of position 
in a galaxy. Earlier studies by Rubin 
and Ford (1971) indicated the presence 
of ionized gas (Fig. 8) in the nuclear 
bulge extending out to 2 kpc from the 
nucleus of M31. This gas is diffuse in 
spatial distribution and the measured 
radial velocities of the Ha and [N II] 
6583 A lines have been used by Ford and 
Rubin for kinematical investigations. 
Here we report on the intensities of the 
lines emitted by this gas and explore the 

The principal emission lines that are 
seen in the spectra are the [0 II] 3727- 
29 A doublet, [0 III] 5007 A, Ha, the 
[N II] 6548-83 A doublet, and the [S 
II] 6717-31 A doublet. A spectrum of 
the nuclear bulge of M31, taken with 
the Carnegie image tube plus the Kitt 
Peak 4-m spectrograph is shown in Fig. 
8. The lines are weak (e.g., W(6563 A) 
» 1 A), and we have measured the 

H« [Nil] [SlO 


NE 400 pc 



Fig. 8. Spectrum along major axis in nuclear bulge of M31, showing emission lines from 
diffuse gas superimposed upon continuum produced by integrated starlight. Carnegie image 
tube plus 4-m Cassegrain spectrograph, plus N 2 -baked Hla-J plates, exposure 120 min; original 
dispersion 28 A/mm. Absorption lines crossing plate at right are night sky absorption lines 
from0 2 (1-0 band). 



intensities of these lines between 50 pc 
and 100 pc distance from the nucleus 
only. The intensities of the lines cor- 
rected for the underlying stellar absorp- 
tion in the case of H« and [S II] 6717 A 
(due to Ca I) and for interstellar extinc- 
tion with E (B — V) = 0". l ll are given 



[0 II] 3727-29 


[N II] 6548-84 


[0 III] 5007 


[O I] 6300 


[S II] 6717-31 




The intensity ratio of the [S II] lines 
6717/6731, as well as earlier measure- 
ments by Munch (1960) of the 3727/ 
3729 ratio of [O II], indicates low elec- 
tron densities (^ 50 cm -3 ). The ob- 
served line intensities cannot be attrib- 
uted to conventional time-independent 
H II regions, particularly because of the 
large strength of the [S II] lines. On the 
other hand, the emission spectrum bears 
a strong resemblance to the spectra of 
supernova remnants (SNR) such as the 
Cygnus loop. Because the velocity dis- 
persion in the central regions is about 
150-200 km s -1 , one might expect cloud- 
cloud collisions to simulate the spectrum 
of the Cygnus loop, which is expanding 
at 140 km s -1 . However, in-galactic SNR 
the [O III] 5007 A line is weaker than 
the [O II] 3726-29 A doublet by about 
a factor of 3, whereas in M31 we find 
the opposite. Furthermore, an excess ul- 
traviolet flux (A < 2500 A) has been 
observed in M31 by the orbiting astro- 
nomical observatory. There is sufficient 
theoretical as well as observational justi- 
fication to attribute this UV flux to stars 
that are about to become white dwarfs. 
The observed UV excess in M31 is suffi- 
cient to influence the ionization equi- 
librium of the gas we are studying, and 
hence simple cloud collisions cannot be 
the sole agents for ionizing and exciting 
the interstellar gas. One can hope to 
learn about the evolution of the hot (Te 

«* 60,000°K) pre- white dwarf stars 
from observations of the ionized gas. 
Already our observations rule out the 
possibility that these stars shine for 
about 10 7 years at high temperatures. If 
these stars lose most of their energy by 
neutrino emission, as has been suggested, 
they would live for only 10 5 years, in 
which case one has to use a time- 
dependent model for the ionization of the 
gas by these hot stars. The spectral line 
intensities we have measured are con- 
sistent with this model. A qualitative 
prediction of this model is that at in- 
creasing distance from the nucleus the 
[O III] 5007/H* line ratio should de- 
crease, possibly accompanied by an in- 
crease in the [0 I] 6300/H« line ratio. 
It is to be noted that the cloud collision 
model makes no such prediction. We plan 
to secure further observations to check 
this and carry out time-dependent model 
calculations if the data warrant them. 


Munch, G., Expanding motions in inter- 
stellar gas in the nuclear region of 
Messier 31, Astrophys. J., 131, 250, 

Rubin, V. C, and W. K. Ford, Jr., 
Radial velocities and line strengths of 
emission lines across the nuclear disk 
of M31, Astrophys. J., 170, 25, 1971. 


Search for Shell Structure in 

Association with Old Galactic 

Supernova Remnants in Neutral 

Hydrogen and Formaldehyde 

George E. Assousa 

The study of the neutral hydrogen 
distribution in the vicinity of old galactic 
supernova remnants (SNR) is part of 
our effort in both the northern and 
southern hemispheres to understand the 
effects of the SNR on the interstellar 
medium. Thus our observations have led 
to the discovery of expanding fragmen- 



tary shells surrounding the SNR's HB21 
(Assousa and Erkes, 1973), S147 (As- 
sousa, Balick, and Erkes, Year Book 73, 
p. 899), and W44 (Knapp and Kerr, 
1974; Sato, 1974). Examination of the 
neutral hydrogen distribution from the 
Assousa, Balick, Erkes SNR survey 
(Year Book 73, p. 899) has also shown 
evidence for similar partial shell distri- 
butions in the vicinity of the Cygnus 
Loop and W41. 

The discovery of the shell structure in 
the neutral hydrogen prompted us, in 
collaboration with Carl Heiles of the 
University of California, Berkeley, and 
Robert Sanders from the National Radio 
Astronomy Observatory, to examine the 
SNR's HB21 and S147 for the presence 
of H 2 CO to determine the likelihood of 
molecular formation in these cool ac- 
creted shells. It was hoped that the 
detection of formaldehyde in these 
clouds would: (1) test the hypothesis 
that molecules can form behind isother- 

mal shocks and therefore could be asso- 
ciated with regions in which hydrody- 
namical disturbances have occurred. 
This could, for example, be an explana- 
tion of dense molecular line regions near 
the galactic center; (2) allow detailed 
kinematic mapping of expanding SNR 

The H.CO observations were carried 
out at 6 cm with the 85- foot antenna at 
the Hat Creek Radio Observatory. The 
beamwidth is 10' at 6 cm, and beam 
efficiency is 0.79. A cooled parametric 
amplifier gave a system noise tempera- 
ture of 60°K and the 100-channel filter 
bank was used with 30 kHz filters giving 
a total window of 3 MHz or a velocity 
coverage of 187 km/sec. Peak-to-peak 
noise for the 30 kHz filters is 0.04° to 

A total of 144 positions for S147 and 
126 for HB21 were covered in 0.33° grids 
in I and b except near the edges of HB21, 
where in directions perpendicular to the 

TABLE 1. List of Supernova Remnants Observed for H 2 CO 







(arc minutes) 

Detection of H 2 CO 









































































NGC 7822 

























VRO 42.05.01 















0607 + 17 












edge of the remnant the spacing was 10 
arc minutes. No absorption line could be 
observed in either set of profiles. 

To further test possible association of 
H 2 CO with SNR's, 23 additional galactic 
remnants were observed. All objects are 
listed in Table 1. Twelve remnants 
showed possible to definite lines. Of these, 
five large remnants were observed in de- 

tail, CTB72, W63, NGC 7822, HB3, and 
PKS0607+17. Two, NGC 7822 (Fig. 9), 
and HB3 (Fig. 10), showed absorption 
lines at several positions along the rem- 
nant. For these sources, the depth of the 
absorption line varies with position, but 
there is no apparent change in the line's 
central velocity. 
From this we conclude that: (1) be- 

NGC 7822 



'sr = - 86.5 km/sec 

+ 20 +5 -10 -25 -40 -55 

Velocity km/s 

-70 -85 


Fig. 9. H 2 CO absorption spectrum in the direction of NGC 7822. The primary feature is 
5.6 km/sec wide and appears at Vi ar = —14.9 km/sec. A weaker feature appears at vi sr = 
—86.5 km/sec. 



Fig. 10. H 2 CO absorption spectrum in the direction of HB3. A prominent feature, 7.4 km/sec 
wide, appears at V\ 8T = —40.5 km/sec. 


cause there is no velocity variation, the (SNR's) (Erkes, Year Book 68, p. 368; 
absorption could not be taking place in Assousa et al., Year Book 73, p. 899) 
any possible expanding shell associated offers an interesting possibility, however, 
with the remnant; (2) based on the de- Theories of the evolution of SNR's pre- 
tailed observations of the large remnants diet that the general interstellar mag- 
there appears to be no association be- netic field may be considerably amplified 
tween the H 2 CO observed and any of as the SNR compresses the medium sur- 
the remnants; (3) the variation in depth rounding it. We therefore chose two 
of the absorption features observed indi- promising candidates, HB21 and W44, 
cates either structure in the continuum and attempted to detect the Zeeman 
source or the absorbing clouds. We are splitting of the 21-cm hydrogen line in 
currently studying these absorption fea- the shells surrounding them. The com- 
tures with the hope that we can deter- ponent of the magnetic field along the 
mine the nature of the excitation (Evans line of sight will produce opposite fre- 
et al., 1975). quency shifts of the radiation of opposite 

circular polarization. The total shift is 

References very small, only 2.8 Hz/microgauss, so 

a very narrow line must be observed for 
Assousa, G. E., and J. W. Erkes, An very long integration times, with con- 
expanding shell of neutral hydrogen siderable care taken to avoid instru- 
surrounding the supernova remnant mental effects. 

HB 21, Astron. J., 78, 885, 1973. In collaboration with Linda K. De- 
Evans II, N. J., B. Zuckerman, G. Noyer from the University of Illinois, 
Morris, and T. Sato, Interstellar H 2 CO Urbana, and Joseph W. Erkes from Dud- 
I. Absorption studies, dark clouds, and ley Observatory, Albany, observations 
the cosmic background radiation, were made on the 140-foot telescope of 
Astrophys. J., 196, 433, 1975. the National Radio Astronomy Observa- 
Knapp, G. R., and F. J. Kerr, A cold HI tory with a specially designed "Zeeman" 
shell around the supernova remnant feed which switches between the two 
W44, Astron. Astrophys., 33, 463, 1974. opposite circular polarizations. No field 
Sato, F., Neutral hydrogen associated was detected in the HB21 shell. Using a 
with W44, Publ. Astron. Soc. Jpn., 26, three standard deviation confidence limit, 
459, 1974. we may say J5 she n < 12 microgauss. Ap- 
plying a theoretical amplification factor 
of 4, we can then estimate that the gen- 
Zeeman Splitting of the 21-cm Line eral interstellar magnetic field in this 
of Neutral Hydrogen in Supernova region must be less than 3 microgauss. 
Remnant Shells The situation in the direction of W44 

.is much more complicated, as shown in 
Fig. 11. A theoretical curve correspond- 

The problem of measuring the inter- in S to 5 microgauss is shown fitted to a 

stellar magnetic field has been an ex- portion of the spectrum. The fit is not 

tremely difficult one. Indirect methods unique, however, and this observation is 

lead us to infer that it is quite small, sti11 under stud y- 
on the order of a few microgauss, but no 

really satisfactory direct determinations d„, „ „ 

. / . J . . . . . . Keference 

exist, except in a few special objects, 

which do not permit us to obtain a value Verschuur, G. L., Further measurements 

for the general field (Verschuur, 1970). of the Zeeman effect at 21-centimeters 

The discovery of expanding HI shells and their limitations, Astrophys. J., 

around some supernova remnants 161, 867, 1970. 

Kenneth C. Turner 





ZZ' 91 '0 ct'0 80 "0 ti0*0 00 "0 h0 ' Q 

(») 3dnibd3dN31 UNN31NU 

80 '0 

si •&- 























& 1 CO 
3 CO 
CO ^3 




+3 P 







O U3 


+3 -M 



S ^ 


T3 o3 









fl .2 


•43 £ 



N &D 



i — i 


2 03 






wo circul 


-^> CD 


of the 
um corr 




an spec 


a) a 


tical Zee 


2 ° 

fafi +3 

PR c3 



Magellanic Stream Gas 
Kenneth C. Turner 

The problem of the history and dy- 
namics of the Magellanic Clouds is an 
interesting and complicated one. These 
two galaxies, usually classified as irreg- 
ulars, are our two nearest neighbors in 
space, being only some 60 kpc from our 
own Galaxy. Indeed, they are sufficiently 
near that our Galaxy exerts an order of 
magnitude more gravitational force on 
each cloud than they do on each other. 
We would therefore expect this to be 
a very disturbed system, and such is 
indeed the case. Not only is the velocity 
field of each cloud quite complicated 
(being, in general, multiple valued), but 
a considerable amount of neutral hydro- 
gen exists between the two clouds (J. V. 
Hindman et al., 1963, and Year Book 67, 
p. 290, and 68, p. 366). Arguments have 
been made attempting to relate this ma- 
terial to possible histories of the Galaxy- 
Magellanic Cloud System (see K. C. 
Turner et al., Year Book 67, p. 290, and 
references cited in I. Mirabel and K. C. 
Turner, 1973). 

The discovery of the Magellanic 
Stream (Mathewson et al., 1974), a very 
long stream of neutral hydrogen extend- 
ing from the general direction of the 
Magellanic Clouds to well past the South 
Galactic Pole, obviously has a very im- 
portant bearing on these arguments, and 
it was thought important to confirm this 
discovery and to obtain profiles of this 
material for detailed dynamical analysis. 

Figure 12 shows spectra observed in 
the direction of the brightest concentra- 
tions near the Magellanic Clouds. Here 
the material is readily detectable, having 
velocities not far from those associated 
with the Clouds themselves. Farther 
away, near the South Galactic Pole, 
however, the material has radial veloc- 
ities near those of our local galactic gas. 
However, our wide velocity resolution, 
chosen for high sensitivity, together with 

some base-line problems to be discussed 
in another section of this Report, made 
it impossible to make a clear division 
between Local and Stream hydrogen. 


Hindman, J. V., R. X. McGee, A. W. L. 

Carter, E. C. J. Holmes, and M. 

Beard, A low resolution hydrogen-line 

survey of the Magellanic system, Aust. 

J. Phys., 16, no. 4, 552-569, 1963. 
Mathewson, D. S., M. N. Cleary, and 

J. D. Murray, The Magellanic Stream, 

Astrophys J., 190, 291, 1974. 
Mirabel, I., and K. C. Turner, A search 

for neutral hydrogen remnants of 

strong tidal disruption of the Small 

Magellanic Cloud, Astron. Astrophys., 

22, 437, 1973. 

A Search for Very High 
Velocity Clouds 

Merle A. Tuve, Charles A. Little, 
and Everett T. Ecklund 

A cumulative search has continued 
through the year for hydrogen clouds of 
very high velocities of approach, which 
might be clouds from outer space " fall- 
ing" into our Galaxy, or, conversely, 
clouds of very high velocities of reces- 
sion, if they are explosively expelled in 
some way from, for example, the center 
of our Galaxy. As indicated in the Re- 
port for last year, a routine was estab- 
lished for searching selected sky points 
from zero velocity to — 800 km/sec and 
from zero to +800 km/sec. A 53-channel 
receiver is used, modified to give each 
channel a velocity width of 16 km/sec 
(to half-power points) and spaced by 
16 km/sec. The 60-foot parabolic an- 
tenna has a half-power width of 0.87° 
on the sky. 

Observations are made at even 10° 
intervals in galactic longitude from 0° 
through 350°. At each longitude, obser- 
vations are made at galactic latitudes 
00°, 04°, 08°, 12°, 20°, 24°, 28°, 40°, 









L = 278, B = -54 


-350 -250 -ISO -S3 50 150 




Fig. 12. Relative spectra of Magellanic Stream neutral hydrogen (cold sky reference spec- 
trum has been subtracted). Deviations from zero near zero velocity represent changes in the 
amount of local hydrogen between the stream point and the reference point. Stream gas 
velocity is near 175 km/sec in the local standard of rest. 

Charles J. Peterson 


48°, 52°, 60°, and at the same list of change, cause changes in the base-line 

negative latitudes. Thus, 21 sets of ob- level. 

servations are made for each galactic It thus appears from our modest sur- 
longitude for which the listed points rise vey that hydrogen clouds from outer 
above the horizon. Of 756 possible points, space falling toward our Galaxy are not 
about 580 rise above the horizon at this a conspicuously frequent occurrence in 
latitude, and about 540 of these had been space, and clouds expelled from our Gal- 
observed by May 1975. For each sky axy at high velocities are also hard to 
point four runs are made, two receding find, if they exist, 
and two approaching. Each run is the 
average of three independent integra- 
tions of 80 seconds each on each channel. The Structure of Globular Clusters 
The overall receiver noise temperature 
is a bit high (about 300°K) in aver- 
age runs, and the tuning curve of the Among the most spectacular objects of 
parametric amplifier and other receiver our Galaxy are the 120 globular clusters, 
circuits varies some from day to day (or near i y spherical systems each of which 
hour to hour, if adjustments are made), contains up to 10 6 stars. The globular 
From critical inspection one concludes c l us ters were formed early in the history 
that a diffuse cloud of great velocity of the Galaxy; their distribution in space 
spread (say 200 or 400 km/sec) would an( j their kinematical properties hold 
fail to be identified if its emission raised c i ues to the processes involved in the 
the antenna temperature only %° or formation of the Galaxy as a whole. 
1°K; it would simply appear to be a Tne se systems, because of their appar- 
modest change in curvature of the base ent i y s i mp i e structure, also provide a 
line. However, a cloud of moderately bas } c tes t for dynamical theory. Theo- 
wide velocity spread (150 km/sec or retical studies (King, 1966) show that 
less) would show as a distinct bulge in a duster may be defined in terms of three 
the (curved) base line if it contributed parameters: a total mass, a core radius, 
~ 3°K to the antenna temperature. and a limiting radius. The core radius is 
Careful examination of hundreds of t he radius at which the surface density 
records indicates that a cloud of small of stars has dropped to one-half that of 
velocity spread, ~ 50 km/sec or less, t he central value; it is closely related to 
would clearly be noticed if it gave rise the potential energy with which the 
to an antenna temperature of %°K or c i us ter formed (e.g., a small core radius 
more - indicates a tightly bound cluster). The 
To date, no confirmed example of such limiting radius defines the "edge" of the 
a cloud has been found. A list of about c i us ter. It is not an intrinsic property of 
20 suspected examples remains for re- the c i uster but is set by the external 
examination among the many small devi- grav itational field of the Galaxy at that 
ations in the routinely observed curves, point f the cluster's orbit where the 
which have been studied and repeated. It ga l a ctic tidal forces are strongest. This 
should also be noted that a small accu- point is at the closest approach of the 
mulation of neutral hydrogen on one cluster to the center of the Galaxy and 
side of the Galaxy, swept up by galaxy j s ca n e d the perigalacticon distance, 
motion through a very low density of Of particular interest are the very 
neutral hydrogen in outer space (if such distant clusters which, because of their 
exists) would not be detected if it gave uncertain relation to the Galaxy at the 
rise to an antenna temperature of only time of their discovery, were designated 
a few degrees, because other effects, such "intergalactic tramps." Until recently, 
as antenna "spillover" (maximum 8°K) these distant clusters had been little 



studied because of the faintness of their 
individual stars. One of these objects, 
Palomar 2, is of particular interest be- 
cause it is observed in the anticenter of 
our Galaxy, and hence a study of its 
orbit, stellar composition, and colors may 
teach us also about the outer regions of 
our Galaxy. Pal 2 has an estimated dis- 
tance modulus of 19 or 20 (McCarthy 
and Treanor, 1964) and thus is between 
60 and 100 kpc from the sun or 70 to 
110 kpc from the center of the Galaxy. 
Pal 2 therefore is located at least twice 

as far from the center as the generally 
adopted radius of our Galaxy (which is 
of the order of 30 kpc; the sun is about 
10 kpc from the center). 

As part of a long-term study of the 
properties of globular clusters (Peterson, 
1974; Peterson and King, 1975) and as a 
start of a study of Pal 2, star counts 
have been made on two plates of this 
cluster, one plate taken with the 84-inch 
(Lynds) and one at the prime-focus of 
the 4-meter telescope (Rubin and Ford). 
Approximately 600 cluster stars are 


r c 

i i — 






O 4M- 



t f ° 

D E-83 





- 'o. 









* ***** * * 



; * 1Q-; * , 






1 ...... J 

i i i i i i_ 








Fig. 13. Surface density versus radius from star counts of two plates (E-83, 84" Kitt .feak 
telescope; 4M-1465, 4-meter Mayall telescope) in the distant globular cluster Palomar 2. The 
dashed line is a theoretical model curve from King (1966) with core radius log r c ' = —1.08; 
tidal radius log r% = 0.67; and concentration parameter c = log (r t '/r/) = 1.75. The insert 
is a copy of a 4-m prime focus plate of Palomar 2, taken on N 2 -baked 0-127 plate, exposure 
70 min ; original scale 12"/mm. 



countable within a radius of 5' on a one- 
hour exposure at the Mayall telescope. 
The surface density profile (Fig. 13) can 
be fitted quite well by a theoretical pro- 
file computed from a self-consistent dy- 
namical model (King, 1966) for a con- 
centration class c = log (r c /r t ) — 1.75, 
where r c is the core radius and r t is the 
limiting or tidal radius. The latter radius 
is relatively insensitive to model curves 
and is r t = 4'7 ± (T2 (for c = 1.75) 
corresponding to a tidal radius of r t = 
80-140 parsecs at the estimated distance 
of the cluster. From the tidal radius we 
can determine the closest approach of 
the cluster to the center of the Galaxy. 
This distance is much smaller (by at 
least a factor of 5) than the presently 
observed distance for Pal 2. Hence Pal 
2 must orbit the Galaxy in a greatly 
elongated orbit, which carries it to vast 
distances from the Galactic Center. 


King, I. R., The structure of star clusters. 

III. Some simple dynamical models, 

Astron. J., 71, 64, 1966. 
McCarthy, M. F., and P. J. Treanor, 

Infrared observations of Abell 2, Ric. 

Specttrosc. Lab. Astrosfis. Specola 

Vaticana, 6, 511, 1964. 
Peterson, C. J. Distribution of orbital 

eccentricities of the globular clusters, 

Astrophys. J. (Lett.), 190, L17, 1974. 
Peterson, C. J., and I. R. King, Observed 

radii and structural parameters in 

globular clusters, Astron. J., 80, 427, 



Image Tube Systems 

W. Kent Ford, Jr. 

During the past two years, greatly 
improved versions of the Carnegie image 
tube have become available for astro- 
nomical observations. These cascaded 
image intensifies, RCA type C33063 

EP3, are in large part a result of a two- 
year development program undertaken 
at RCA for the Institution and supported 
by a grant to the Department of Ter- 
restrial Magnetism from the National 
Science Foundation. With the comple- 
tion of this development effort, our em- 
phasis has shifted somewhat from the 
testing and evaluation of prototype tubes 
to the design and construction of opti- 
mum systems incorporating the improved 
tubes and to observing with them. An 
excellent opportunity for this was pro- 
vided by my joint Carnegie- AURA ap- 
pointment to Kitt Peak National Observ- 
atory as a Visiting Resident Scientist. 

The gain of an image intensifier is 
derived from the quantum efficiency of 
the photocathode that converts the opti- 
cal input to an electronic image. Various 
image tubes store and detect the elec- 
trons which contain the pictorial infor- 
mation by different means. In the cas- 
caded intensifier, primary photoelectrons 
are accelerated through 10 to 15 kilo- 
volts and focused on an efficient electron 
multiplier consisting of a phosphor- 
photocathode sandwich. The secondary 
photoelectrons that are thus created are 
in turn accelerated and focused on a 
phosphor screen which then displays an 
intensified optical image. This intensi- 
fied image is recorded on a photographic 
emulsion with a transfer lens. 

In a high-gain image intensifier, each 
primary photoelectron produces an opti- 
cal scintillation consisting of tens of 
thousands of photons of which only a 
few percent are collected by the transfer 
lens. Nevertheless, these photoelectron 
scintillations are sufficiently bright to be 
recorded as a clump of blackened grains 
in the developed photographic plate; the 
certainty of detection of any individual 
scintillation is proportional to its bright- 
ness. Image intensifies with a radiant 
power gain of 100,000X or greater, when 
used with a fast transfer lens, produce 
individually detectable scintillation im- 
ages and hence are particularly useful 



in recording an image consisting of a 
very few photoelectron events in the 
absence of any background. Such high- 
gain systems have also been used to 
great advantage in faint object spectros- 
copy by Lynds and others at Kitt Peak 
by moving the photographic plate per- 
pendicular to the dispersion to provide 
widening of the spectrum. Without this 
plate-widening mechanism the bright 
scintillations in the spectrum of a stellar 
object tend to overexpose or saturate the 
plate with relatively few photoelectrons, 
and consequently only high contrast 

10 sec F 1.0 

features of the spectrum are detected. 
The moving-plate scheme is particularly 
useful in obtaining widened spectra with 
a high signal-to-noise ratio of faint point 
source objects as it permits convenient 
use of a very small entrance aperture, 
which is necessary in order to discrimi- 
nate against the night sky background. 
The technique, however, is limited to 
point source objects, since any spatial 
information is lost in the widening 

In order to obtain high signal-to-noise 
spectra of extended objects having low 

20 sec F 1.4 

40 sec F 2.0 

80 sec F 2.8 

160 sec F 4.0 

Fig. 14. Test sequence with image intensifier illustrating method of improving signal-to- 
noise ratio in exposures to approximately the same optical density by averaging more photo- 
electron scintillations. 


surface brightness, some other tech- tion problems associated with nonlinear 
nique is required. A two-stage cascaded photographic process. To circumvent 
intensifier such as the RCA C33063 these problems we have continued our 
operating at 28 or 30 kilovolts has suffi- work on the electrical readout of intensi- 
cient gain (radiant power gain of 15,000 fiers with an image dissector along the 
to 20,000 X) that with an f/1 transfer lines described in some detail in Year 
lens the statistical fluctuation or quan- Book 71 (p. 221). The method is based 
turn noise due to the accumulation of on the fact that photoelectron scintilla- 
individual photoelectrons is recorded tions as displayed at the output of a 
even on a fine-grain emulsion. This sys- high-gain intensifier have a lifetime 
tern is fast but noisy, since each photo- associated with the persistence of the 
electron scintillation is statistically con- phosphor screens. By sampling each pic- 
tributing many blackened grains to the ture element in a line image with an im- 
exposure. The plate appears very grainy age dissector (an electrically scanned 
in uniformly exposed portions, and weak photomultiplier) with a repetition rate 
features in the image are lost in the large comparable to the decay time, a histo- 
density fluctuations. To improve the gram of the occurrence of photoelectron 
signal-to-noise ratio, more photoelectron events can be accumulated in a digital 
scintillations must be integrated in a signal averager. 

longer exposure. If this is done with the The work thus far may be summarized 
fast transfer lens, the plate will be ex- as follows: The image dissector, lens 
cessively dense; it is usually more expe- coupled to the two-stage image intensi- 
dient simply to stop-down the transfer fier, operates with the efficiency predicted 
lens so that the optical density of the for the particular dissector photocathode 
recorded image is near optimum for the quantum efficiency. This efficiency is 
particular emulsion. adequate for repetitive single-line scans 
This situation is illustrated by a series and is probably useful for two-line scans 
of test exposures reproduced in Fig. 14 of spectra of star-plus-sky and sky. The 
and by the various examples of long- efficiency is below that of an ideal pho- 
slit galaxy spectra in other portions of toelectron counter because of the decay 
this Report. The test spectra are of a characteristic of the phosphors. The 
continuum source (a tungsten lamp) system is satisfactorily free of fixed pat- 
mixed with an emission source (helium, tern noise due to the lens rather than 
argon, and neon discharge lamps). The fiber optic coupling of the intensifier and 
exposures are with a two-stage Carnegie dissector. The present data acquisition 
image tube, RCA C33063 EP3 S/N system operates in an increment memory 
8'7'2*1, and are recorded on an N 2 baked by one mode for each detected photo- 
IllaJ plate through a Repro-Nikkor 85 electron scintillation rather than the add 
mm f/1 transfer lens. On this plate the scaler count to memory contents mode 
transfer lens is changed by one stop and that has been very successfully used at 
the time doubled for each exposure. The Lick Observatory by Robinson and 
improvement in the detectability of weak Wampler. We have shown that the sta- 
lines is evident. tistical noise associated with the recorded 

data is just the noise associated with 

The Intensifier-Image the number of counts per channel. 

Dissector Scanner The intensifier-image dissector scan- 
ner appears to be one of the best of the 

W. Kent Ford, Jr. available digital detectors. Because of 

The accuracy of measurement of in- the persistence of the phosphors that are 

tensities is limited in the photographic used as the storage element, the effec- 

readout of image intensifies by calibra- tiveness of the device is limited to only 



a few lines or possibly a small- format 
area array. The advantages of our par- 
ticular device are the relative freedom 
of fixed-pattern noise, good resolution, 
and good statistical counting properties. 

The Pointing of Our 60-ft 
Derwood Radio Telescope 

E. T. Ecklund, N. Thonnard, and K. C. Turner 

With the ever increasing amount of 
radio interference we have been experi- 
encing at low frequencies at our radio 
astronomy observatory in Derwood, 
Maryland, it has been clear for some 
time that our observational future there 
lies at short wavelengths (~ 15 mm) 
where there are few sources of inter- 
ference and where only very nearby 
sources will be troublesome due to the 
line-of-sight propagation of these short 
wavelengths. We have developed a good 
receiver at these frequencies and have 
undertaken various observational pro- 
grams. As reported in Year Book 73 (p. 
919), we have readjusted the surface of 
the antenna and more than doubled its 
efficiency. But, at such short wavelengths, 
a 60-foot antenna has a half-power 
beam- width of only 0.058°, and our 
pointing has been poor enough that so 

far we have only been able to study 
sources that could be seen in "real time." 
Before investing a lot of time and 
money to install more accurate position 
indicators on our antenna, we wanted to 
determine where the weak links in the 
pointing were and how rigid the feed 
support and paraboloid were. Strain 
gauges consisting of music wire and a 
special scale readable from any attitude 
were wired from the edge of the parabo- 
loid to the focus and vertex and across 
both sides of the paraboloid. On a cloudy, 
still day, the telescope was moved in all 
directions and with a small telescope the 
reading of the strain gauges was taken. 
This enabled us to calculate the actual 
position of the focal pqint relative to the 
paraboloid at any attitude. It was grati- 
fying to discover that the antenna itself 
and the feed support are very rigid. In 
Fig. 15 are plotted the measured deflec- 
tions of the focus relative to the parabo- 
loid versus the angle from the zenith the 
antenna is pointing to. The deflections 
are only a function of the zenith angle 
and can be fitted to the equation (solid 

A0 = O°O187sin(Z) 

which can be easily introduced as a sim- 
ple correction. Measurements made of 

0.020 - 

£> 0.015 - 

f> 0.010- 

0.005 - 

1 1 1 1 

• Motions in meridian plane 



1 1 


x Motions east of meridian 
O Motions west of meridian 

o« ^___ 








s^ x 




1 1 











Z (degrees) 

Fig. 15. Angular deflection of the feed support relative to the focal point of the 60-ft 
Derwood paraboloid for different antenna positions as a function of angular distance from the 
zenith Z. The s'olid curve has been fitted to the data and gives the deflection as A0 = 0.0187 
sin Z. 



the edge of the dish relative to the polar 
axis indicate that the deflections are also 
less than 0.02% in this plane, indicating 
that most of the pointing problems we 
have been encountering are probably due 
to large deflections in the large driving 
gear and in the position encoders them- 
selves, which are coupled to the driving 
gear. If after a few more measurements 
we can confirm that there are no large 
and unusual deflections between the 
polar and declination axes and the para- 
boloid itself, then we feel that by placing 
an accurate angle encoder on the polar 
and declination axes we will be able to 
utilize the antenna to its full capability 
at short wavelengths. 

A New Receiver for the IAR 

N. Thonnard and K. C. Turner 

Because of the serious stability and 
base-line problems experienced while 
using the present receiver at the Instituto 
Argentino de Radioastronomia to study 
very faint signals, and as our scientific 
interests are moving toward research 
that will require much improved sensi- 
tivity and stability, it has been decided 
to design and construct a completely new 
system that will take advantage of the 
many technological advances made since 
the development of our original hydro- 
gen-line receiver. 

Our present system may be thought of 
as being divided into three parts. The 
front end, consisting of parametric am- 
plifier, switch, feed horn, mixer, tempera- 
ture control apparatus, etc., all installed 
at the focus of the telescope, will be 
replaced by a completely new, modern 
system. Dimensional and thermal sta- 
bility, as well as low noise, will be em- 
phasized to ensure that the system does 
not vary with telescope position and time. 

The intermediate frequency part, con- 
sisting of IF amplifiers, 2nd mixers, and 
filter banks, we propose to leave un- 
changed. The back end, which converts 
the output energy of the filters to nu- 
merical values, will be completely rede- 

signed. In place of the present system of 
tuned amplifiers, diode phase detectors, 
and large integrating capacitors with a 
mechanical readout switch, the proposed 
system will sample the slightly inte- 
grated output of the filters directly, per- 
forming the synchronous detection and 
integration digitally in a small on-line 
computer, which will also continuously 
measure the gain of the system. This will 
eliminate many additional possibilities 
of instability and instrumental drift and 
increase the flexibility of observing pro- 

A joint proposal, by the DTM and 
the Instituto Argentino de Radioastro- 
nomia, to the National Science Founda- 
tion and the Ajencia Executivo Argentino 
has been made to perform this moderni- 
zation, and to make observations of 
southern spiral galaxies and high-velocity 
clouds as an international effort under 
the joint agreement for scientific co- 
operation signed by Argentina and the 


This year again, the astronomers at 
DTM acknowledge with thanks the gen- 
erous allotments of telescope time from 
Kitt Peak National Observatory, Na- 
tional Radio Astronomy Observatory, 
and Hat Creek Radio Observatory. 

The development of the high-gain 
image tube and the work on the digital 
system have been supported in large part 
by a grant from the National Science 


Ever since the early 1960's when the 
Instituto Argentino de Radioastronomia 
was established in a cooperative venture 
between the Carnegie Institution of 
Washington, the Argentine Consejo Na- 
cional de Investigaciones Cientificas e 
Tecnicas, the Comision de Investiga- 
ciones Cientificas de la Provincia de 
Buenos Aires, and the Universities of La 
Plata and Buenos Aires, relations be- 



tween our institutions have been warm 
and cordial, weathering the problems 
that plague any international coopera- 
tive venture, due primarily to the con- 
fidence that between friends and col- 
leagues difficulties can always be worked 
out. That this continuing collaboration 
bears fruit in both hemispheres is dem- 
onstrated by this brief outline of IAR 
activities, as well as the research de- 
scribed in more detail elsewhere in this 


During 1974 several minor changes in 
the 21-cm line receiver were made. A 
new parametric amplifier pump system 
was installed, using a Gunn diode, which 
results in a more reliable operation of 
the front end. The low- frequency section 
of the local oscillator was redesigned and 
built with integrated circuits. It will be 
used to drive the Fairchild phase-lock 
oscillator. Several power sources were 
replaced by solid-state units. Some work 
was done with a new front end using the 
same type of paramp as the one actually 
in use. In the last months of the year, 
the new installation for the receiver room 
was worked on. Several temperature 
controllers were designed and built to 
replace those in the front end and filter 
banks. Work was also done on the design 
of low-noise 30 MHz preamplifiers and 
100 KHz band-width IF amplifiers. 

21-cm Work 

F. R. Colomb, M. Gil, and R. Morras 
have extended their observations around 
the south celestial pole. They have fin- 
ished the observations and final reduc- 
tion of the region 290° ^ I ^ 314°; 
-17° ^ b ^ -27°; with Al = Ab — 
1°. In these surveys they found several 
intermediate negative velocity clouds 
(IVC). The survey is now in the final 
stage of preparation for publication. 

F. R. Colomb, W. Poppel, and C. 
Heiles of the University of California 
continued the observations for the hy- 
drogen survey at low velocities (—50 

^= v ^ -f 50 km/sec) of the southern 
hemisphere. They completed about 30% 
of the observations. 

F. R. Colomb, R. Morras, and W. G. 
L. Poppel made an attempt to detect 
Comet Kohoutek (1973) in the con- 
tinuum and in the line, with negative 

New Personnel 

E. Bajaja returned from Leiden, The 
Netherlands, where he was working on 
H I line observations of galaxies with 
the new Westerbork synthesis radio- 
telescope. T. Gergely returned from 
Maryland, U.S.A., where he received his 
Ph.D.; he is now studying supernova 
remnants. M. Caponi returned from 
Maryland, U.S.A., where she received 
her Ph.D.; she is studying theoretical 
aspects of galactic structure. 

Two fellowships were awarded by the 
Provincial Research Council to Mar- 
garita Franco and Carlos Alberto Olano, 
who have begun investigations at the 
Institute under the supervision of Drs. 
F. R. Colomb, E. Bajaja, W. Poppel, 
and I. Mirabel. Three new technicians, 
Ing. A. J. Sanz, Ing. J. J. Mignacco, and 
Enrique Eduardo Hurrell are now en- 
gaged in the design and maintenance of 
our equipment, under the supervision of 
E. M. Filloy. 

H. O. Morganti began work in the 
photography laboratory preparing draw- 
ings for publications. S. Acero was en- 
gaged for the operation of the radio 

I. F. Mirabel and M. Franco have 
finished the study of a neutral hydrogen 
structure at I = 8°, b = —5.5° with a 
radial velocity of —218 km/sec. This 
feature was probably ejected from the 
galactic nucleus. 

I. F. Mirabel, W. G. L. Poppel, and 
E. R. Vieira completed the study of an 
object discovered by Cugnon near I = 
349°, b = +3°. The results will be 
published in Astrophysics and Space 

W. G. L. Poppel and E. R. Vieira of 


the Institute* de Fisica, of the Universi- Z-dependence of all the experiments 

dade Federal de Rio Grande do Sul, known to us as well as the shape of the 

Brazil, completed the observations and spectrum of the heavy-ion continuum, 

reductions of their large survey for the Our formula is by no means the complete 

region 240° — I — 372°, + 3° ^ b explanation, but it has made successful 

^ -j-17° '. The first part of their atlas of predictions. It points toward a different, 

profiles and contour maps (360° ^ I heretofore untried mechanism. 

f= 372°) has been published (Carnegie Investigations with the beam of polar- 

Inst. Wash. Pub. No. 633). The rest will ized protons are typical of research hav- 

follow. Using the atlas, a study of the ing an advanced theoretical framework 

southern extension of a near cloud ob- on which to build. Theory and experi- 

served by R. Sancisi and H. van Woerden ment are used together to determine the 

(1970) was started by C. A. Olano and structure of nuclei, and the work is slow 

W. G. L. Poppel, as well as an analysis and tedious. The experiments are often 

of the general motions of gas in the in energy regions of isotopic structure 

region 348° ^ I ^ 372° by M. Franco that have seen little activity in years, 

and W. G. L. Poppel. regions that yield information only to 

Observations were made to detect neu- new measuring techniques. During the 

tral hydrogen in three galactic clusters past year we have probed the structure 

by W. G. L. Poppel and M. D. Vota and of two fundamental isotopes, 10 B and 

in six small dark clouds by M. Gordon 16 0. In one we suspected no anomalies 

and W. G. L. Poppel. and found one; in the other we suspected 

one and found another. 

Sancisi, R., and H. van Woerden, An Continuum and Characteristic 

elongated neutral-hydrogen emission X Rays Produced with Potassium Ions 

feature in Scorpius and Ophiuchus, L. Br own, N.Thonnard, and C.K.Kumar 
Astron. Astrophys., 5, 135, 1970. 

The creation of vacancies in the elec- 
tron shells of atoms by the impact of 
point charges is fairly well understood. 

L. Brown, G. H. Pepper, N. Thonnard, The probability that an electron, a pro- 

C. K.Kumar, and L.G. Arnold ton Qr Qne of the Ught nudei ^ eject 

Introduction an electron from the K or L shell can be 

_ n calculated for a wide range of velocity 

L. Brown ■. , , • j j i 

and atomic number, provided any elec- 

Two completely different modes of re- trons of the projectile are much less 

search originated from work with the tightly bound than those of the target. 

Van de Graaff machine during the past These calculations, based on the Born 

year. Investigations with heavy ion- and the binary encounter approxima- 

induced x rays are in the confused state tions, fail when applied to collisions in 

typical of a new field of study. For vari- ■ which the electron shells of target and 

ous reasons most experiments by others projectile interact. Theoretical attempts 

have been conducted at energies well to describe collisions of this kind gen- 

beyond the capabilities of our machine, erally start with Fano and Lichten's 

e.g., 200 MeV Kr, and our examination proposal (1965) that the wave functions 

of the low-energy region showed an of the separated atoms transform through 

aspect of this phenomenon so extreme molecular orbitals into the wave func- 

and so simple that it seemed to cry for tions of an atom having a nucleus made 

explanation. This led us to a theoretical of the target and projectile nuclei in 

clue: an equation that reproduces the close proximity. This way of picturing 




collisions has explained the results of 
several experiments but has not led to a 
quantitative theory of vacancy produc- 
tion with Z t ~ Z 2 > 10 (for a review, 
see Garcia et al., 1973). It was for this 
reason that we undertook the experiment 
described in the last Annual Report 
(Year Book 73, p. 929). 

A result that particularly impressed 
us was the exponential dependence on 
target atomic number, Z 2 , of the emission 
of potassium K lines. A change in yield 
of nearly three decades was shown in 
Fig. 43 of Year Book 73 as Z 2 took on 
values from 21 to 30; the atomic number 
of the projectile is 19. At the time no 
theory could account for this; an adap- 
tation of the molecular orbital model 
has since been proposed by Meyerhof 
et al. (private communication), but it 
is based on an empirical equation and 
fails to reproduce the strong Z depend- 
ence of our results. We have developed 
a formula that does reproduce the rela- 
tive Z dependence of the data just cited 
as well as the results of all similar 
experiments known to us. 

In the lower right of Fig. 16 is shown 
schematically how the energies of the 
wave functions of the separated atoms 
K and Sc (right) transform through 
molecular wave functions to form briefly 
a united atom with Z = 19 + 21 (left). 
If the wave functions make this trans- 
formation adiabatically, the electrons of 
the two Is shells of the heavy and light 
atom will form, respectively, the Is and 
2p shells of the united atom when the 
two nuclei are in close proximity and 
reform themselves on separation. This 
predicts no K-shell vacancies and hence 
no x-ray emission. Experiments showed 
early that the light collision partner 
emits copious amounts of characteristic 
x-rays, the heavy partner relatively few. 
This has been taken as evidence that the 
2po- molecular orbital is ionized, and 
various ideas have been proposed for 
doing this but none has yielded quanti- 
tative results. 

Figure 16 also shows the energies of 


i r (pm) 5 

(ICT ,8 sec) 2 

U.A. a /Z 
_g Closest approach 

Fig. 16. Energy levels of the diatomic system 
K + Sc. In the lower right is the correlation 
diagram for the inner shells. The united atom 
structure is at the left; the separated atom 
structures are at the right with Sc denoted by 
dashed lines and K by solid lines. The re- 
mainder of the figure shows the energies of lso- 
and 2po- molecular orbitals as functions of the 
internuclear separation, which has been con- 
verted to units of time through the ion velocity. 

the lso- and 2po- molecular orbitals as 
functions of internuclear distance. The 
energy is also shown in units of electron 
oscillation frequency and the distance of 
separation in units of time obtained 
through the ion velocity. This demon- 
strates that the transformations of the 
lso- cannot be adiabatic because the elec- 
tron executes only a few oscillations as 
r decreases below 5 pm. The conse- 
quences of this violation cannot be stated 
with certainty, but it is reasonable to 
expect electrons trying to follow the lso- 
orbital to undergo quantum jumps, which 
will frequently (perhaps always) be 
radiationless transitions that eject from 
the 2p<r or higher orbitals. This method 
of ejecting electrons in atomic collisions 
is the novel part of our proposal. 



The collision times are short compared 
with the times of spontaneous electro- 
magnetic transitions, so such a process 
must be induced if it is to be effective. 
We have assumed the stimulating field 
to result from the motion of the projec- 
tile nucleus acting on electrons whose 
average position is given by the expec- 

tation value of the united atom radius. 
The Fourier transform of this time- 
varying field gives the frequency spec- 
trum of the virtual photons with which 
the electron may interact. Assuming that 
Auger transitions in the lso- orbital are 
stimulated in proportion to the ampli- 
tude of this spectrum yields a formula 




-2 2 


^I0" 26 










5 MeV CI 

4 7 MeV 



45 MeV Ni xlO 

-J> „ 

2 MeV K 

° J50 MeV 


40 Zo 60 

80 100 

Fig. 17. Cross sections for K-vacancy production as, function of Z 2 for eight experiments. 
The value given is for the emission by the light collision partner. Solid and open circles dis- 
tinguish nearby data. The projectile and incident energy are indicated for each group. Vertical 
arrows indicate Zi. The 2 MeV K points are from the experiment described in Year Book 73, 
p. 929; the 5 MeV CI are from Winters et al. (1975) ; the 45 MeV Ni are from Kubo et al. 
(1973); the 150 MeV Xe are from Gippner et al. (1974a); the remainder are from Meyerhof 
et al. (private communication). The dashed curves are from an empirical equation by Meyer- 
hof et al. (1974) ; the full curves are from Equation 1. 



for the relative probability I(Z 2 ) of 
ejecting an electron from the 2po- orbital 

I(Z 2 ) oc exp 


3 e 2 Zx + Z, 




where e is the electronic charge; H, the 
reduced Planck constant; and v, the 
projectile velocity. 

The characteristic x-rays of these col- 
lisions are superimposed on a continuum 

whose origin is not yet certain. One pro- 
posal, discussed in Year Book 73, p. 929, 
suggests that the continuum results from 
transitions to vacancies in the molecular 
orbitals. The energy of the orbital is a 
function of the internuclear distance, as 
illustrated in Fig. 16, so a continuum of 
photon energies results. Recent work by 
Greenberg et al., 1974, and Thoe et al. 
(1975) indicates that this continuum is 

6 - 

10* - 

io 4 - 

I0 3 - 

icr - 

o - 

E ph (keV) 

Fig. 18. X-ray continuum yield per unit energy interval in arbitrary units as a function of 
photon energy. Vertical arrows indicate the united atom limit. The top three measurements 
are from Meyerhof et al. (1974) ; the fourth, which has pulse-height analyzer points repro- 
duced by a curve and a shaded region, is from Gippner et al. (19746) ; and the bottom, which 
has a different (1 to 3 keV) energy scale from the other four, is from the experiment described 
in Year Book 73, p. 929. The straight lines are from Equation 2 with n — 2 for the 2 MeV K 
on V, and n = 1 f or the others. 



stimulated. The time-varying field we 
assumed to induce the radiationless 
transitions should also induce transitions 
in the molecular orbitals, if they have 
vacancies to be filled. This assumption 
yields a formula for the relative intensity 
1(E) of the continuum as a function of 
photon energy E 

1(E) oc exp 


3n 2 



2 Z, + Z 2 Uv) (2) 


where a is the Bohr radius and n the 
principal quantum number. Continua 
have been observed in what seem to be 
two bands: one that should terminate on 
the K shell of the united atom, and the 
other on the L shell. The former uses 
n ■== 1 in Equation 2; the latter, n = 2. 

Figure 17 shows a wide range of char- 
acteristic x-ray cross-section measure- 
ments resulting from heavy ion collisions. 
The solid lines show the application of 
Equation 1; the dashed curves are from 
the empirical equation of Meyerhof et 
at., private communication. Equation 1 
was derived when only the 2 MeV K+ 
and the 47 MeV 1+ data were at hand; 
the remaining measurements allowed its 
predictive capabilities to be tested. Fig- 
ure 18 shows measurements of x-ray 
continua with solid lines from Equation 
2. As the comparisons in Figs. 17 and 
18 show, the two equations account for 
relative values of the Z dependence of 
characteristic emission and the E de- 
pendence of continuum emission rather 
well. One cannot calculate cross sections 
with them at present, or the dependence 
of the data on the projectile energy, 
although the relative dependence is sat- 
isfactory at all energies tested. 

These two formulae result from the 
same starting assumption and appear to 
link together two different observations 
of the same collision. The anomalous 
nature of these collisions was recognized 
more than seven years ago, and the ab- 
sence of any other quantitative explana- 
tion forces further study of the two 
formulae. The interpretation we attach 
to Equation 1 — that of stimulated Auger 

transitions — is new to atomic physics 
and has potentially a wider applicability 
in atomic collisions than just the pro- 
duction of K-shell vacancies. 


Fano, U., and W. Lichten, Interpretation 
of Ar+-Ar collisions at 50 keV, Phys. 
Rev. Lett., 14, 627, 1965. 

Garcia, J. D., R. J. Fortner, and T. M. 
Kavanagh, Inner-shell vacancy pro- 
duction in ion-atom collisions, Rev. 
Mod. Phys., 45, 111, 1973. 

Gippner, P., K. H. Kaun, W. Neubert, 
F. Stary, and W. Schulze, Excitation 
of K X-rays by bombardment of thick 
solid targets with 150 MeV Xe ions, 
Joint Institute for Nuclear Research, 
E7-7688, Dubna, 1973a. 

Gippner, P., K. H. Kaun, F. Stary, W. 
Schulze, and Yu. P. Tretyakov, Quasi- 
molecular K X-ray excitation by bom- 
bardment of Ge atoms with Ge ions, 
Nucl. Phys., A230, 509, 19746. 

Greenberg, J. S., C. K. Davis, and P. 
Vincent, Evidence for quasimolecular 
K X-ray emission in heavy ion colli- 
sions from the observation of the X- 
ray directional anisotropy, Phys. Rev. 
Lett, 33, 473, 1974. 

Kubo, H., F. C. Jundt, and K. H. Purser, 
Target Z dependence of projectile K 
X-ray production cross sections in 
high energy, heavy ion-atom collisions, 
Phys. Rev. Lett, 31, 674, 1973. 

Meyerhof, W. E., T. K. Saylor, S. M. 
Lazarus, A. Little, B. B. Triplet, L. F. 
Chase, Jr., and R. Anholt, Molecular- 
orbital K X-ray formation in heavy 
ion collisions, Phys. Rev. Lett., 32, 
1279, 1974. 

Meyerhof, W. E., T. K. Saylor, S. M. 
Lazarus, A. Little, B. B. Triplett, R. 
Anholt, L. F. Chase, Jr., and P. D. 
Bond, K vacancy production in heavy 
ion collisions, private communication. 

Thoe, R. S., I. A. Sellin, M. D. Brown, 
J. P. Forester, P. M. Griffin, D. J. 
Pegg, and R. S. Peterson, Observation 
of large and strongly energy dependent 



directional anisotropics in nonchar- 
acteristic K x rays emitted in heavy 
ion collisions, Phys. Rev. Lett., 34, 
64, 1975. 
Winters, L., M. D. Brown, L. D. Ells- 
worth, T. Chiao, E, W. Pettus and J. 
R. Macdonald, K X-ray production 
in single collisions of chlorine and 
sulfur ions, Phys. Rev., All, 174, 1975. 

The 15 N(p,«) 12 C Reaction Excited 
with Polarized Protons 

G. H. Pepper, L. Brown, and L. G. Arnold 

In the isotope 16 0, a group of five 
levels located in the energy region from 
12.3 to 13.3 MeV excitation energy 
stands out as pure compound nuclear 
states when observed through the reac- 
tion 15 N(p,a) 12 C. This region includes the 
isobaric analogs of the ground and first 
three excited states of 16 N and 16 F. At 
low energies the cross section is small be- 
cause of the Coulomb barrier; at higher 
energies it is very small for about 1.5 
MeV because no states are excited. Ref- 
erence to the limited energy level dia- 
gram of Fig. 19 (from Ajzenberg-Selove, 
1971) would seem to belie these remarks, 
because it shows more than five states, 
but 15 N(p,«) 12 C can proceed only 
through intermediate states that are 0+, 
1~, 2+, 3~, 4+, • • • , a consequence of 
the conservation of angular momentum 
and parity. States with other values of 
J* are invisible to it. 

In addition to furnishing information 
about the structure of 16 this reaction 
is also the last one in the carbon-nitrogen 
cycle through which main-sequence stars 
with temperatures hotter than our sun 
burn hydrogen. Details of the reaction 
in the proton energy range of a few 
hundred keV are important for predict- 
ing the reaction rates at thermonuclear 
energies. The resonance at 0.34 MeV is 
very near this energy range and the four 
near 1 MeV cannot be neglected. 

Three questions about the structure of 
16 in this region need answers. First, 
the location of the 2+ state at 13.1 MeV 

is uncertain by ±0.15 MeV; there may 
in fact be two such states. Second, there 
is evidence from other light nuclei that 
some states of given J* have orbi- 
tal angular momentum admixtures of 
Al = 2. The ground state of the 
deuteron was the first example of this; 
Year Book 73, p. 920, describes just such 
an occurrence in the structure of 4 He. 
The presence of a d-wave admixture 
(Al = 2) in the two 1~~ states would 
alter the manner of extrapolating these 
resonances to low energies. Third, there 
is reason to question whether the assign- 
ment T = and T = 1 to the pair of 3~ 
states at 13.13 and 13.26 MeV is valid. 
The T = 1 would be the analog of the 
second excited state of 16 N and 16 F. The 
13.13 MeV state shows up strongly in 
12 C(a,«) 12 C and weakly in 15 N(p,«) 12 C; 
the 13.26 MeV state is just the opposite 
— behavior that is normal for such T 
assignments — but the relative magni- 
tudes and their y-decays to T = states 
bring this into question and cause one 
to suspect that the analog state is the 
pair, not just one of them. Some states 
in the mass 8 system, which were the 
subject of previous investigation (see 
Year Book 70, p. 333, and Year Book 71, 
p. 247), exhibit similar features. Excit- 
ing the reaction with polarized protons 
should answer the first two questions, 
and once we are certain of the structure, 
the matter of the isospin of the 3~ states 
can be studied through calculations that 
couple the two through the isospin term 
of the nuclear potential. 

We measured six angular distributions 
of polarization analyzing power for pro- 
ton energies from 0.34 to 1.21 MeV, all 
at energies where differential cross- 
section data were available. The results 
are described through an expansion of 
Legendre functions (see Year Book 73, 
p. 921, for the defining equations). The 
experimentally determined coefficients Ai 
of the associated functions, which de- 
scribe the analyzing power, are shown in 
Fig. 20. Their rapid change as functions 
of proton energy complicates the analy- 



E x J*.T 



















.- . 2^0 










- 0.338 


N + p 

Fig. 19. Limited energy level diagram for 16 and isobar diagram for A = 16. Levels, as they 
were thought to exist prior to our experiment, are indicated by horizontal lines with the 
energy in MeV above the ground state at the left and two quantum numbers descriptive of 
the state at the right — the angular-momentum-parity and the isospin. The energies of 16 N 
have been shifted vertically to eliminate the neutron-proton mass differences and the Coulomb 
energy. Levels that are presumed to be isospin multiplets are connected by dashed lines. A 
scale at the right gives proton energy for ^NCp^^C in MeV. Only levels with J 71 " = + , 1", 
2 + , 3", . . . can be observed with this reaction and T = 1 states will show up much less 
strongly than T = states. 



0.4 - 

0.2 - 


-0.2 - 

E p (MeV) 

Fig. 20. Coefficients of the associated Legendre functions that describe the measured analyz- 
ing power of 15 N(p,a) 12 C. These coefficients can be calculated theoretically, if the structure of 
the compound nucleus, in this case 16 0, is known. The location of resonances accepted at the 
time this work began are indicated with crosses; the vertical line is the resonance energy; the 
horizontal gives the width. 

sis, so as a further aid we measured the 
analyzing power in small energy steps 
at the angles where the functions P\(Q) 

and P\(0) are zero. The analyzing power 

for the 0.34 MeV resonance, the 12.44 
MeV 1~ state, is evidence that it has a 
small d-wave admixture. The 1 _ state 
at 13.09 MeV shows pronounced polar- 
ization effects, which can be explained 
as interference from the nearby 3~ and 
2+ states as well as a small d-wave ad- 
mixture. The 3~ at 13.26 MeV shows 
polarization effects resulting from its in- 
terference with l - and 2+ states. Theo- 
retical calculations of the coefficients, 
which should establish the structure ac- 
curately, are under way. 


Ajzenberg-Selove, F., Energy levels of 
light nuclei A = 16-17, Nucl. Phys., 
A166, 1, 1971. 

The 9 Be(p,n) 9 B Reaction Excited 
with Polarized Protons 

L. Brown, U. Rohrer, and L. G. Arnold 

In Year Book 72, p. 191, we described 
our study of 10 B with 9 Be(p,p) 9 Be using 
polarized protons. For proton energies 
below 2 MeV (8.4 MeV excitation energy, 
see Fig. 10, p. 192, Year Book 72) we 
were able to demonstrate an unambigu- 
ous and simple structure. Above 2 MeV 
proton energy there were differential 



cross-section data at only one angle and 
this lack prevented the phase-shift 
analysis from giving as clear a picture 
of the structure there. At 2.56 MeV pro- 
ton energy (8.89 MeV excitation energy) 
lie two excited states, degenerate in 
energy within experimental error. Marion 
(1956) showed that one, observed through 
9 Be(p,«y) 6 Li, was 2 + and the other, ob- 
served through 9 Be(p,n) 9 B, was 3+. 
Later Altman et al. (1962) concluded 
after examining the relative neutron and 
proton reduced widths that the latter 
state must be 3~. 

Our phase shift analysis clearly fav- 
ored the 2+ and 3+ at 2.56 MeV and 
was unable to fit the data with 2+ and 
3~. We have completed the measurement 
of the angular distribution of the analyz- 
ing power of 9 Be(p,n) 9 B at four energies 
that span the structure in question as 
well as two energies well above it. The 
coefficients of the associated Legendre 
functions are shown in Fig. 21 and give a 




2 + 



o- — -o v 




^x* A3 



^ A 2 _ 



• A, 






E p (MeV) 


Fig. 21. Coefficients of the associated Legen- 
dre functions that describe the analyzing 
power of 9 Be(p,n) e B. At 2.56 MeV proton en- 
ergy are two states degenerate in energy, a 2 + 
of width 38 keV and a 3 + or 3" of width 85 keV. 
These data rule out the 3" because the inter- 
ference of two states of the same parity leads 
to Ai = Az = over the range of the reso- 

clear answer to the question of parity 
assignment. If the two states have the 
same parity, the odd coefficients will be 
zero across the 85 keV wide resonance, 
a condition not satisfied. In addition, a 
nonzero A^ coefficient requires the pres- 
ence of d waves in both the proton and 
neutron channels. The 3~ assignment is 
favored. The shell model predicts a 3~ 
state in 10 B near 8.89 MeV excitation 
energy that has a nearly pure d 5/2 con- 
figuration. The absence of an A 4 co- 
efficient is consistent with a 3~ state that 
is a pure configuration. 

The results of the 9 Be(p,n) 9 B analysis 
require us to examine again the results 
of our earlier study of 9 Be(p,p) 9 Be near 
2.56 MeV. By using the shell model pre- 
dictions for the 3~~ state as a guide, we 
hope to clarify the matter. 


Altman, A., W. M. MacDonald, and J. B. 

Marion, Nucleon reduced widths and 

charge independence, Nucl. Phys., 35, 

85, 1962. 
Marion, J. B., Excited states in 10 B, 

Phys. Rev., 103, 713, 1956. 

A Return to 4 He (p,p) 4 He 

L. Brown and G. H. Pepper 

In Year Book 73, p. 927, we described 
the present status of 4 He(p,p) 4 He, which 
was the first experiment we undertook 
with the polarized proton beam (see 
Year Book 65, p. 72). The phase shifts 
describing this interaction are given by 
various parameterizations, each capable 
of reproducing all measurements over a 
20 MeV range with accuracies around 
1%. We use this interaction from time 
to time to verify the polarization of our 
beam, and this time continued it for a 
few weeks at the request of R. L. Walter 
of Duke University who hoped he would 
be able to distinguish between the vari- 
ous parameterizations. The matter has a 
special significance because measure- 
ments in our energy range should be able 



to decide the existence of a d-wave ad- 
mixture to the ground state of 4 He for 
which Plattner et al. (1975) have ad- 
vanced evidence, evidence resting pri- 
marily on our old data. The measure- 
ments we just completed tend to favor 
one of the parameterizations, but a clear 
answer requires accuracies five times 
better than we generally run, which 
means running 25 times longer. This ex- 
periment will probably be postponed 
until the current of our polarized beam 
is correspondingly greater, or until we 
are convinced the results justify the ex- 
penditure of time. 


Plattner, G. R., R. D. Viollier, and K. 
Alder, Evidence for the D State of 
4 He, Phys. Rev. Lett., 34, 830, 1975. 


In December G. H. Pepper partici- 
pated in measurements of K- and L-shell 
x-ray cross sections produced with 7-35 
MeV lithium ions at the Tandem Ac- 
celerator Laboratory of Florida State 
University. This work was undertaken 
to test theories for creating inner-shell 
vacancies through direct interaction, i.e., 
where the projectile may be considered 
a point charge. Analysis is not complete 
but the K-shell data seem to indicate 
a need for relativistic wave functions in 

high-Z targets when struck by low- 
velocity projectiles. 

During September A. J. Ratkowski of 
New York University used the alkali 
beam of the DTM Van de Graaff to 
measure the energy dissipated by elec- 
tronic collisions when these projectiles 
were stopped in silicon crystals. The 
targets were silicon surface-barrier de- 
tectors that registered inherently the 
electronic excitation through the separa- 
tion of electron-hole pairs. The results 
follow the theory of Lindhard et al. 
(1963 a, b, 1968) for the stopping of 
slow, heavy atomic projectiles in matter. 


Lindhard, J., M. Scharff, and H. E. 
Schiott, Range concepts and heavy 
ion ranges, (Notes on Atomic Colli- 
sions, II), Mat. Fys. Medd. Dan. Vid. 
Selsk., 33, No. 14, 1963a. 

Lindhard, J., V. Nielsen, M. Scharff, and 
P. V. Thomsen, Integral equations 
governing radiation effects (Notes on 
Atomic Collisions, III), Mat. Fys. 
Medd. Dan. Vid. Selsk, 33, No. 10, 

Lindhard, J., V. Nielsen, and M. Scharff, 
Approximation method in classical 
scattering by screened Coulomb fields, 
(Notes on Atomic Collisions, I), Mat. 
Fys. Medd. Dan, Vid. Selsk., 36, No. 
10, 1968. 


T. I. Bonner, D. B. Cowie, B. H. Hoyer, N. R. Rice, and R. B. Roberts 


As in the past several years, much of 
the work of the Biophysics Section has 
dealt with the DNA of eukaryotic cells. 
Bonner is attempting to isolate DNA 
segments which are transcribable in vivo 
and which therefore represent essentially 
all of the gene regions. He is exploiting 
the observation that RNAs from several 
species have been found to contain short 
(~ 25 nucleotides) transcribed runs of 

poly (adenylic acid). The complemen- 
tary DNA molecules can in principle be 
isolated simply by taking that fraction 
of total denatured DNA which binds to 
a column of poly (A) or poly(U), to iso- 
late the partner DNA strand. 

Hoyer is continuing his study of re- 
latedness among the primates and re- 
ports on experiments measuring simi- 
larity of single-copy DNA among several 
rather closely related African monkeys. 



Repeated DNA has also been examined, 
and a preliminary finding of a family 
present in human but not in chimp DNA 
is reported. 

Repeated sequences in rat and mouse 
DNAs are the subject of Rice's report. 
She has found a small class (about 1% 
of total DNA) of repeated mouse (or 
rat) DNA which reassociates to higher 
thermal stability with homologous DNA 
rather than with heterologous DNA. 
How these two species can each possess 
rather homogeneous families which are 
only loosely related to each other is an 
evolutionary puzzle of great interest. 

Cowie, working at the Pasteur Insti- 
tute with Georges Cohen and Paolo 
Truffa-Bachi, has continued his study of 
the Escherichia coli allosteric enzyme 
aspartokinase I-homoserine dehydroge- 
nase I. Work has centered on the con- 
formational changes induced in the 
enzyme when it binds to specific anti- 
bodies and on the effects of the enzyme's 
ligands on this antibody coupling. 

Finally, in the hope that it will be of 
general interest, we include Roberts's 
account of the remarkable history of 
the Biophysics Section — from the early 
days when radioisotopes for biological 
experiments were produced by DTM's 
cyclotron, through periods concentrating 
on metabolic pools, ribosomes, and DNA. 
The group is now disbanding, and we 
offer this as tribute to the section of 
which we are pleased and privileged to 
have been members. 

Properties of Human DNA Sequences 
Containing Poly (dA) or Poly (dT) 

T. I. Bonner 

The messenger RNA of the cellular 
slime mold (Dictyostelium discoideum) 
contains a short sequence of poly (adeny- 
lic acid) of about 25 nucleotides in addi- 
tion to the longer poly (A) sequence (in 
this case about 100 nucleotides) which 
is commonly found at the 3' end of 
eukaryotic messenger RNAs (Jacobson 
et al, 1974). In contrast to the long 
poly (A) sequence which is added to the 

precursor nuclear RNA after it is tran- 
scribed, the short poly (A) is transcribed 
from DNA. Furthermore, in at least half 
of the messenger RNA molecules, the 
short poly (A) is separated by no more 
than 25 nucleotides from the long 
poly (A) at the 3' end of the molecule. 
This result suggests that the short 
poly (A) sequence functions as a termi- 
nator of the transcriptional unit. If the 
presence of short poly (A) sequences with 
similar properties were common through- 
out the eukaryotes, it could prove to be 
a very useful tool in the study of gene 

There is evidence that these sequences 
are indeed common. Bishop et al. (1974) 
have studied the amount of poly(dA) 
present in the genomes of several euka- 
ryotes. While they found substantial 
quantities of poly(dA), they concluded 
that there was much too little poly(dA) 
of sufficient length for the long poly (A) 
in messenger RNAs to be encoded by the 
DNA. However, if one studies their re- 
sults it is apparent that the genomic 
poly (A) reacts with poly(U) to form 
hybrids which melt approximately 20° C 
lower than hybrids of synthetic poly (A) 
with poly(U). Assuming that the lower- 
ing of the melting temperature is due to 
the short length of the genomic poly (A) 
and not to the presence of other bases 
within the poly (A), the length must be 
about 20-30 nucleotides. With an aver- 
age length of 25 nucleotides, the amount 
of poly (A) in rat DNA implies about 
140,000 poly (A) sequences per genome 
or one poly(A) 25 sequence per approxi- 
mately 20,000 nucleotide pairs. 

It has also been shown (Nakazato et 
al, 1974) that the heterogeneous nuclear 
RNA of a human cell line contains about 
one poly (A) sequence of 20-40 nucleo- 
tides per molecule. Furthermore, the 
short poly (A) sequences are clearly the 
result of transcription, while the long 
poly (A) sequences are not. However, 
they do not find any short poly (A) in 
cytoplasmic RNA, an observation which 
is not consistent with the short poly (A) 


being near the end of the message unless short, their thermal stabilities when re- 

the long poly (A) is formed by elonga- acted with long molecules of poly (A) or 

tion of the short poly (A). poly(U) will be quite low. For this rea- 

The discussion above clearly suggests son one must do the reactions under 
that the occurrence of large numbers of higher salt and lower temperature con- 
20-30 nucleotide poly (A) sequences is a ditions than usual. However, going too 
common feature of eukaryotic DNAs. far in this direction would allow non- 
It also appears likely that these se- specific reactions to take place. The 
quences occur approximately once per conditions we have used, namely 0.28 M 
transcriptional unit, although it is not PB at 25°C, provide a reasonable corn- 
clear whether they are usually located promise. Human DNA of 2300 nucleo- 
within the messenger sequences. The iso- tides median length binds 7%-14% 
lation of DNA sequences containing the compared to an expected value of 6%, 
short poly (A) stretches and the comple- assuming that human DNA contains the 
mentary sequences containing short same amount of short poly (A) as rat 
poly(T) stretches could be useful for and that the poly (A) sequences are not 
several reasons. First, the short poly (T)- clustered. However, Escherichia coli 
containing sequences should be enriched DNA, which contains about 30 times less 
in expressable sequences, thus allowing poly (A) (Bishop et al., 1974), binds 
more sensitive measurements of differ- about 20 times less DNA to the column 
ences in expression between different cell even in the presence of human DNA 
populations. Second, the isolated se- bound to the column. This result shows 
quences should represent only one strand that the reaction criterion is not so low 
of the DNA and should thus be incapa- as to allow the nonspecific binding of the 
ble of self-reaction. This property alone E. coli DNA directly to the column or 
would be a great technical convenience, even secondarily to the bound human 
Third, the isolated sequences could be DNA. 

useful, especially if the short poly (A) I have used two different binding pro- 
sequence is located at the 3' end of the cedures. In the first procedure the DNA 
message, in the mapping of the sequence was alkali denatured, then brought to 
organization of the genome using the 0.28 M PB and loaded on the poly (A) - 
poly (A) sequence as a reference point, agarose, or poly (U) -agarose, column so 

In principle the isolation can be ac- that the sample volume was entirely 
complished simply by reacting denatured within the column volume. After 10-15 
DNA with synthetic poly (A) or poly (U) minutes the column was washed with 
that is covalently bound to a matrix more 0.28 M PB, and the bound DNA 
such as agarose. Thus, the sequences was eluted with solutions containing at 
containing poly(dT) can be bound to a least 70% formamide and less than 0.09 
poly (A) -agarose column. The bound M PB. The second procedure used a 
sequences should include those from column of 1 cm diameter by 3 cm height 
which short, poly (A) -containing RNA of either poly(U) or poly (A) -agarose 
molecules are transcribed. Unfortunately, through which the sample was passed at 
the bound fraction may also include a rate of about 1 ml/min. The first pro- 
some sequences that are representative cedure gives as much as twice as much 
of the other DNA strand, since it is binding as the second. However, when 
known (Molloy et al, 1972) that there labeled poly(dA) of 38 nucleotides aver- 
are also short U-rich regions in RNA. age length or poly (dT) of 48 nucleotides 

average length is bound using the second 

Column Properties procedure, 96% binds to the column. 

Since the poly (A) and poly(T) Thus, the increase in binding using the 

stretches in the DNA are expected to be first procedure must be due to the re- 



association of molecules that cannot 
directly bind to the column with those 
that can. Since the only substantial dif- 
ference between the two procedures is 
approximately 10 minutes of incubation 
time, the reassociation must involve se- 
quences that are highly repetitive under 
our very relaxed incubation conditions. 
This explanation is supported by the 
observation that of the DNA which 
binds to a poly(A)-agarose column in 
15 minutes only 55%-65% will bind a 
second time, while a third binding gives 
80%-90%. This pattern of binding indi- 
cates that rapidly reassociating sequences 
not containing poly(T) are being suc- 
cessively removed. Thus two or three 
binding steps are necessary to purify the 
poly (A)- or poly (T) -containing se- 
quences to a reasonable extent. As we 
shall see later, most of the molecules 
that bind to the column contain highly 
repeated sequences. 

Since the poly (A) and poly(U) that 
are covalently bound to the agarose are 
nominally about 1500 nucleotides long, 
which is considerably longer than the 

longest poly (A) stretches in natural 
RNAs, the whole length of a poly (A) or 
poly(U) sequence in a bound RNA or 
DNA molecule should be base-paired. 
And since the stability of the hybrid 
depends on the base-paired length, we 
should be able to differentiate the bound 
sequences on the basis of the poly (A) or 
poly(U) length by using a gradient elu- 
tion from the column. To test the column 
properties in this regard we have used 
32 P-labeled RNA from HeLa cells, which 
should provide us with at least two size 
classes of poly (A) sequences, about 25 
nucleotides and about 200 nucleotides. 
When this RNA was bound to a poly (U)- 
agarose column, the total binding was 
nearly 2% and the elution pattern typi- 
cally was as shown in Fig. 22. There is 
a pronounced peak at about 65% com- 
pletion of the gradient which must cor- 
respond to the long poly (A) sequences. 
The peak at about 11% completion, 
which is somewhat less pronounced than 
usual, corresponds to the short poly (A) 
sequences, as can be seen in Fig. 23 
which shows the elution pattern of DNA 

Fraction number 

Fig. 22. Gradient elution of T-labeled HeLa from a poly (U) -agarose column. ^P-HeLa 
RNA was bound to a 1 cm X 3 cm poly (U) -agarose column in 0.28 M PB, 10 -3 M EDTA at 
25° C by passing the sample through the column at a rate of approximately 1 ml/min. The 
elution of bound RNA was done with a linear gradient which varied from 0.28 M PB, 10~ 3 M 
EDTA (solution A) at the beginning to 90% formamide, 0.01 M PB, 0.01 M EDTA (solution 
B) at the finish. The refractive index of individual fractions was measured to determine the 
percentage of solution B in the fraction. 




— 80 a) 

18 22 26 

Fraction number 

Fig. 23. Gradient elution of human DNA from a poly (U) -agarose column. 3 H-labeled DNA 
was alkali denatured, brought to 0.28 M PB and then bound to and eluted from a poly(U)- 
agarose column as described in Fig. 22. 

sequences binding to the same column. 
In order to determine how well the 
position in the elution gradient char- 
acterizes the poly (A) sequence length, 
fractions were selected from the early, 
middle, and late portions of the gradient 
and were again bound to and eluted from 
the column. As shown in Fig. 24, at least 
75% of the counts that bind elute within 
two fractions of their original position. 
However, the binding of all three frac- 
tions is only about 50%, which suggests 
that there was substantial degradation 


aaaaaKaaaaa r 

20 40 60 80 

Percent completion of gradient 


Fig. 24. Rechromatography of selected gradi- 
ent fractions. Fractions were selected from the 
early, middle, and late portions of gradients 
similar to the gradient of Fig. 22, bound to the 
poly(U) column and eluted again. The hori- 
zontal bars above the peaks indicate the posi- 
tions of the fractions in the initial gradient. 

between the first binding and the second 
binding. The important point is that 
neither early nor middle fractions have 
an appreciable amount of RNA eluting 
in the late portion of the gradient. This 
means that they have negligible amounts 
of long poly (A) -containing molecules. 
Hence, essentially all of the long poly (A) 
sequences must be base-paired over their 
entire length. The only likely reason that 
elution position would not be directly 
related to the poly (A) sequence length 
is the possibility that base-pairing occa- 
sionally occurs over less than the full 
poly (A) length. Since this does not hap- 
pen to the long poly (A) sequences, one 
can assume that the elution position is 
indeed a good indicator of poly (A) 

The labeled RNA also allows us to 
determine whether the short U-rich re- 
gions in RNA have sufficiently different 
binding or elution properties that they 
can be distinguished from the poly(T) 
sequences in DNA which are comple- 
mentary to the poly (A) 25 sequences of 
RNA. When RNA is bound to a poly (A) - 
agarose column, the midpoint of the 
gradient elution occurs at 21% of com- 
pletion as compared to the midpoint at 
18% completion for the DNA bound to 
poly (U) -agarose shown in Fig. 23. This 
small difference could well be due to 
the increased stability of the rU:rA base 
pairs for the RNA compared to rU:dA 




base pairs with the DNA. Thus it ap- 
pears that we will be unable to dis- 
tinguish between DNA sequences (or 
other complements) corresponding to 
RNA poly(U) and DNA sequences (or 
their complements) complementary to 
RNA poly (A). 

Properties of Poly (A)- or 
Poly (U) -Binding DNAs 

As I mentioned earlier, when DNA is 
bound to the columns a substantial frac- 
tion of the bound DNA will not bind to 
the column a second time. Presumably 
these are DNA molecules containing 
highly repeated sequences complemen- 
tary to molecules that actually bind to 
the column because of their poly (A) or 
poly(U). Two successive bindings to the 
column usually produce a DNA fraction 
capable of binding a third time with 
nearly 90% efficiency. Such fractions, 
which should represent reasonably pure 
poly (A)- or poly (U) -containing se- 
quences, amount to 5-7% of the total 
DNA whose initial median size is 2300 

nucleotide pairs. DNA that has bound 
twice to a poly (A) column, and from 
which the instantaneously reassociating 
fraction (11%) has been removed, re- 
associates with and without an excess 
of unlabeled DNA as shown in Fig. 25. 
In the presence of unlabeled DNA the 
reassociation of the poly (T) -containing 
sequences indicates that 64% of the 
molecules contain repeated sequences 
with an average repetition frequency of 
about 75,000. For comparison, the kinet- 
ics of unfractionated DNA of 2300 NTP 
median length indicates that 46% of the 
unfractionated molecules contain re- 
peated sequences with an average repeti- 
tion frequency of 34,000. Thus the 
poly (T) -containing sequences appear to 
be enriched in highly repeated sequences. 
The kinetics of poly (T) -containing se- 
quences prepared from DNA of 385 
nucleotide average length shows nearly 
as much repeated DNA (about 55%) 
and with nearly as great a repetition 
frequency (about 50,000). Thus, it is 

c t 

Fig. 25. Reassociation kinetics of poly(T)-containing DNA. The reassociation of various 
DNA fractions in 0.14 M PB at 60°C (or the equivalent) are shown as the fraction of labeled 
DNA binding to hydroxy apatite. Reassociation was measured in the presence of a large excess 
of human DNA of 450 NTP average length for the following DNAs: (a) 2300 NTP DNA 
which bound twice to poly(A) columns and was stripped of "instantaneously reassociating" 
DNA, indicated by open circles; (b) unfractionated DNA of 2300 NTP median length, solid 
circles; (c) 385 NTP DNA which bound to poly (A) column, plusses; and (d) poly(T)-primed 
complementary DNA, triangles. In addition, we show the self-reaction of fraction a as open 
squares. The self -reaction Cot values have been multiplied by a factor of 15 to correct for 
sequence enrichment (see text). 



clear that more than half of the poly(T) 
sequences in the genome are within 
about 350 nucleotides of a highly re- 
peated sequence. The estimate of about 
140,000 poly(T) sequences in the rat 
genome includes both poly(T) sequences 
that would occur as poly(U) in RNA 
and poly(T) sequences which would be 
transcribed into poly (A) sequences. 
Thus, the data suggest that perhaps one 
or the other of these two groups is 
adjacent to a single family of repeated 

Figure 25 also shows the kinetics of 
self-reaction of the poly (T) -containing 
sequences. The actual C t values ob- 
tained using the specific activity of the 
DNA have been multiplied by a factor 
of 15 to correct for the fact that the 
sequences, which represent only 5%-7% 
of the genome, are more concentrated 
than they would be if the same amount 
of DNA from the whole genome were 
present. It is quite conceivable that these 
C t values could be in error by as much 
as a factor of 2. With this in mind, the 
data appear to be consistent with the 
interpretation that about 20% of the 
repeated sequences are capable of self- 
reacting with about the same kinetics as 
were evident in the presence of excess 
unlabeled DNA. This result suggests 
that there is still a contamination of 
sequences not containing poly(T) 
amounting to about 10% of the repeated 
sequences or about 6%-7% of the total 
preparation. This level of contamination 
would be consistent with the fact that 
only 88% of this DNA preparation will 
bind to the poly (A) column. 

An initial test to see whether the 
poly (T) -containing sequences are en- 
riched in expressed sequences was per- 
formed by reacting the same preparation 
of DNA with excess RNA from WI-38 
cells to R t = 2200. Even though this 
R t value is an order of magnitude less 
than is necessary to reach saturation 
with total single-copy DNA (Hoyer's 
unpublished results), 13% of the DNA 
molecules react with the RNA. 

Polymerase Copying 

If we take poly(T)- or poly (A) -con- 
taining sequences and use them as tem- 
plate for a DNA polymerase in the 
presence of poly (A) or poly(T) as 
primer, we should be able to generate 
complementary DNA (cDNA) repre- 
sentative of the sequences in one direc- 
tion or the other from the poly (A) or 
poly(T) region. The various possibilities 
are illustrated in Fig. 26. Let A and B 
represent the sequences at the 3' and 5' 
ends of the poly(U) region in an RNA 
molecule and let C and D represent the 
sequences which are at the 3' and 5' 
ends of the short poly (A) region in an 
RNA molecule. Furthermore, let the 
complementary sequences be denoted as 
A', B', C, and D'. Then poly (^-con- 
taining template DNA contains the 
RNA-like sequences A and B and the 
RNA-complementary sequences C" and 
D' as illustrated in the second line of 
Fig. 26. Since the DNA polymerase starts 
at the priming poly (A) sequences and 
proceeds in the 5' to 3' direction, the 
polymerase product will include just the 
sequences B' and C. Of these only the 
sequences B' can react with RNA. Sim- 
ilarly, as shown in the third line of Fig. 
26, poly (A) -containing template DNA 
contains the RNA-like sequences C and 
D and the RNA-complementary se- 
quences A' and B' . Thus, the polymerase 
product contains the sequences A and 
D'. If, as suggested by the slime mold, 
the short poly (A) sequence in RNA 
occurs near the 3' end of messenger 
RNA, then the sequence class D' should 
be complementary to messenger RNA, 
and the sequence class C should be 
largely spacer DNA. 

In an effort to disentangle the proper- 
ties of the sequence classes A, B, C and 
D, I have begun to prepare cDNA from 
poly (A) -binding and poly (U) -binding 
DNAs. I have used the DNA polymerase 
from Micrococcus luteus and have la- 
beled the polymerase product by sup- 
plying 32 P-labeled dGTP to the reaction 
mixture. The use of label in dGTP al- 



(I) RNA Sequences 

A B 
— — uuuuu 


-£ AAAAA — 

. . Poly(A) primed 5 

*£■> polymerase products 

template DNA 


B 1 


RNA complementary 
template DNA 








Poly(T) primed 
polymerase products 





5' 3' 


B" A 




5' 3' 

Fig. 26. Possible DNA polymerase products from poly(A)- or poly (T) -containing templates. 
We denote the sequences at the 3' and 5' ends of poly(U) and short poly(A) sequences which 
occur in RNA by the letters A, B, C, and D as shown in the first line of the figure. The 
complementary sequences are denoted A', B', C", and D' . The remainder of the figure shows 
the various possible types of poly (A)- or poly (U) -binding template sequences and their 
polymerase products. In all cases, AAAAA and TTTTT represent poly(A) and poly(T) 
sequences of about 25 nucleotides. The polymerase products are indicated by dashed lines 
with arrows indicating the direction of copying. 

lows one to measure the incorporation 
of deoxyribonucleoside triphosphates 
into DNA without having to worry about 
the formation of poly (A), poly(T), or 
poly (AT). In the initial attempt using 
(dT) 12 -i 8 to prime poly (U) -binding 
template sequences, incorporation was 
measured by denaturing the polymerase 
product and binding to HAP at 50° C in 
0.042 M PB. The 32 P-labeled material 
that eluted in 0.14 M PB was taken to 
be a single-stranded polymerase product, 
while the material that eluted in 0.42 M 
PB was assumed to be a product con- 
taining a hairpin loop caused by the 
polymerase reversing direction and using 
the previous product as a template. The 
0.42 M PB fraction was discarded, and 
the 0.14 M PB was further purified by 
binding to a poly (A) column. At this 
point there was 47 times as much prod- 
uct in the poly (T) -primed case as in 
the unprimed control reaction. 

Even though the poly (T) -primed 
DNA probably contained a significant 
amount of DNA which is too short to 
form a stable product at 60°C in 0.14 M 
PB, the kinetics of this DNA in the 

presence of an excess of unlabeled DNA 
was studied. As the results in Fig. 25 
show, at least 45% of this DNA contains 
repeated sequences. When reassociated 
with WI-38 RNA to R t = 2000, about 
15% of the DNA reacts with RNA. The 
question whether reacting sequences are 
repeated or are single-copy is of con- 
siderable importance. According to our 
discussion above, the sequences which 
react with RNA must be of the class D' 
which, if the short poly (A) sequences 
are near the 3' end of messenger RNA, 
must be complementary to messenger 
RNA. Since messenger RNA sequences 
are almost entirely single-copy, a large 
fraction of repeated sequences would 
indicate that the short poly (A) sequences 
are not predominately near the 3' end of 
messenger RNA. A very tentative answer 
to this question was obtained by reacting 
the poly (T) -primed DNA with RNA to 
R t = 2600. The product was treated 
with SI nuclease to remove unreacted 
sequences, and the remaining 3% of the 
DNA was reacted with DNA to C t = 
15. The results indicate that at least 
75% of the Sl-resistant DNA contains 



repeated sequences. A more definite 
answer , will require reacting the DNA 
with the RNA to a R t of 20,000-40,000 
so that all the DNA that can react with 
the DNA has an opportunity to do so. 
I intend to pursue the use of the cDNA 
from poly (A)- and poly (T) -containing 
DNA templates in order to determine 
the properties of the adjacent sequences. 
The question of the location of the short 
poly (A) sequences in RNA is of par- 
ticular interest since it bears heavily on 
usefulness of the poly (T) -containing 
sequences as probes for transcriptional 


Bishop, J. O., M. Robash, and D. Evans, 
Polynucleotide sequences in eukaryo- 
tic DNA and RNA that form ribo- 
nuclease-resistant complexes with 
polyuridylic acid, J. Mol. Biol, 85, 
75-86, 1974. 

Jacobson, A., R. A. Firtel, and H. F. 
Lodish, Transcription of polydeoxy- 
thymidylate sequences in the genome 
of the cellular slime mold, Dictyostel- 
ium discoideum, Proc. Nat. Acad. Sci. 
USA, 71, 1607-1611, 1974. 

Molloy, G. R., W. L. Thomas, and J. E. 
Darnell, Occurrence of uridylate-rich 
oligonucleotide regions in heterogene- 
ous nuclear RNA of HeLa cells, Proc. 
Nat. Acad. Sci. USA, 69, 3684-3688, 

Nakazato, H., M. Edmonds, and D. W. 
Kopp, Differential metabolism of large 
and small poly (A) sequences in the 
heterogeneous nuclear RNA of HeLa 
cells, Proc. Nat. Acad. Sci. USA, 71, 
200-204, 1974. 

Relationships of some African 

Monkeys as Determined by their 

Single-Copy DNA Sequence 


Bill H . Hoyer and N. W. van de Velde 

We have examined the relationships 
of several of the African monkeys on the 
basis of the thermostabilities of their 

DNA sequences when single-copy, radio- 
active fragments were reassociated with 
unfractionated, unlabeled DNA frag- 
ments derived from homologous or 
heterologous monkey species. The ra- 
tionale and methodology are essentially 
described in Year Book 69, pp. 488-500, 
and Year Book 71, pp. 260-262. 

These studies provide information on 
the present status of the relationships of 
some African monkeys and provide some 
further insight into their evolutionary 
divergence times. Since these monkeys 
have similar generation times, the rela- 
tive thermostabilities of their reassoci- 
ated sequences may be assumed to have 
a direct relationship to their times of 
divergence (see Year Book 69, pp. 488- 

The radioactive single-copy DNA se- 
quences were derived from tissue culture 
cells from the green monkey (Cerco- 
pithicus sabeus) or the rhesus monkey 
(Macaca mulatta). Single-copy, radio- 
active DNA from the South American 
capuchin monkey (Cebus capucinus) was 
included as a common reference. 

The unlabeled DNA fragments used 
were prepared from the tissues of rhesus 
and green monkeys as well as the capu- 
chin monkey. In addition, DNA frag- 
ments derived from the tissues of the 
baboon (Papio anubis) , the cynomolgus 
monkey (Macaca fascicularis) , and the 
Indochinese bear macaque (Macaca 
arctoides) were used in heterologous re- 
associations. The unlabeled DNA frag- 
ments included the single-copy as well 
as the repeated DNA fragments. 

The results in Table 2 indicate that 
the rhesus monkey is more closely re- 
lated to the other macaques and the 
baboon than is the green monkey. Also, 
the green monkey and the rhesus monkey 
are uniformly diverged from the species 
tested because the ATe 50 '& within each 
group are similar. 

If previous DNA homology data are 
taken into account and the divergence 
times used are realistic (Year Book 69, 
pp. 492-493), it can be inferred that the 



TABLE 2. Divergence of Some African Monkeys as Indicated by the 

Thermostabilities of Radioactive Single-Copy DNA Fragments 

Reassociated with Unlabeled Homologous or 

Heterologous DNAs 

Source of Radioactive DNA Fragments and 
the ATe 5 o* of the Reassociated DNAs 

Source of unlabeled DNA 

C ercopithicus sabeus 

Macaca mv 






C. sabeus 
M. Mulatta 
M. arctoides 
M. fascicularis 
Papio anubis 





* ATe 5 o is determined graphically as described in Year Book 71, pp. 260-261 ; the 
larger the number indicated, the greater the sequence divergence. The radioactive 
DNA fragments were reacted with excess unlabeled DNA ; self-reaction of the 
radioactive DNA was about 1%. The reaction was allowed to proceed to Cot = 
20,000 at 60 °C and 0.14 M PB, and the unique sequences were prepared as indi- 
cated in Year Book 71, pp. 260-261. Escherichia coli 32 P-DNA was included as an 
internal standard. 

African monkeys diverged between 15 
and 30 million years ago. Data derived 
from the radioactive single-copy capu- 
chin DNA sequences indicate a &Te 50 
of 11°C for both green and rhesus mon- 
keys, a result in good agreement with 
previous findings with green monkey in- 
dicator sequences (Year Book 69, pp. 

The African monkeys examined here 
form an evolutionary group somewhat 
more homogeneous than that formed by 
man, orangutan, and gibbon but less 
homogeneous than the group formed by 
man, chimpanzee, and gorilla. 

A Special Class of Related Repeated 
Sequences in Mouse and Rat DNAs 

Nancy Reed Rice 

There are several known cases of re- 
lated satellites occurring within a single 
species or in closely related species 
(Sutton and McCallum, 1972; Rice and 
Straus, 1973; Gall and Atherton, 1974; 
Cooke, 1975). In each case heterogeneity 
within a satellite appears quite limited 
(it has a high T m after reassociation) , 
yet it is clearly dissimilar from its coun- 
terparts (cross product has a low T m ) . 
How two diverging species can each 

possess homogeneous families which are 
only loosely related to each other is an 
unsolved and highly intriguing puzzle. 

Between more distantly related spe- 
cies, where satellites show no cross- 
homology at all, those repeated sequences 
which are related exhibit about the same 
thermal stability when reassociated with 
heterologous as with homologous DNA 
(Rice, 1972). It is as though a particular 
family in each of two diverging lines 
accumulates base changes at about the 
same rate, resulting in the same average 
heterogeneity within each at present. 
Stavnezer (1971), however, has reported 
that there is a fraction of repeated mouse 
(or rat) DNA which reassociates to 
higher thermal stability with homologous 
than with heterologous DNA, thus mim- 
icking the satellite findings. I here report 
results confirming his observation. 

The experiments can be summarized 
as follows: (1) A high thermal stability 
fraction of mouse-repeated 3 H-DNA is 
prepared (mouse satellite having pre- 
viously been removed). (2) 3 H-DNA 
which appears to reassociate instantane- 
ously is removed from this fraction. 
(3) That part of the fraction able to 
reassociate with repeated sequences of 
rat DNA is isolated in double-stranded 



form and treated with SI nuclease, thus 
removing all 3 H-DNA segments which 
are not base paired with the rat DNA. 
(4) Sl-resistant 3 H-DNA is reassociated 
with repeated sequences of rat or mouse 
DNAs and the thermal stability of the 
products determined. 

In one such experiment, mouse L cell 
3 H-DNA from which the satellite had 
been removed following Ag+-Cs 2 S0 4 
centrifugation (as described in Fig. 30 
of the accompanying report) was re- 
duced in piece size to 300 nucleotide 
pairs by sonication, and reassociated to 
Cot 22 in 0.14 M PB at 50°C. Twenty- 
six percent of the total bound to HAP, 
and the 14% of the total which eluted 
above 70° was employed in subsequent 
steps. Instantaneously reassociating ma- 
terial (23%) was removed from the 
latter material by incubation to C t 4 X 
10 -5 and passage through HAP. This 
high thermal stability fraction of mouse 
DNA, which I shall refer to as Fraction 
A, yields a product of predominantly 
high thermal stability when reassociated 
with unlabeled mouse DNA but one of 
considerably lower stability with rat 
DNA {C t 4.5 in each case), as shown 
in Fig. 27. The 63% of Fraction A able 
to reassociate with rat DNA was isolated 
in double-stranded form from HAP (0.4 
M PB elution of bound material), dia- 
lyzed, and treated with SI nuclease for 
2 hours under the high-salt, 37°C con- 
ditions of Hutton and Wetmur (1973). 
Forty-one percent of the radioactivity 
was rendered dialyzable. Nondialyzable 
material (Fraction B) was divided in 
half, reassociated with either mouse or 
rat DNA to C t 16 in 0.14 M PB at 50°, 
applied to HAP, and eluted in 5°C steps 
in 0.14 M PB (Fig. 28). The extent of 
reaction of Fraction B with mouse and 
rat DNAs was very similar (80% and 
74%, respectively), but the two melting 
profiles are quite different (T m with 
mouse DNA about 73°C, with rat DNA 
about 65°C). Thus, it seems clear that 
at least some of Fraction B is able to 
pair with low stability with rat DNA 

and with considerably higher stability 
with mouse DNA. As Fraction B com- 
prises about 4% of total mouse DNA, 
this subclass amounts to at least 1% of 
total DNA. 

The same format was followed in an 


70 80 90 100 

Temperature, °C 

Fig. 27. Reassociation of mouse Fraction A 
3 H-DNA (prepared as described in text) with 
mouse and rat DNAs to Cot 4.5 in 0.14 M PB 
at 50° C. Eighty percent of the 3 H-DNA re- 
associated with mouse DNA and 63% with rat 
DNA; self-reaction totalled 4%. Bound ma- 
terial was eluted from HAP in 5°C steps in 
0.14 M PB, 0.02% SLS. 



Fig. 28. Reassociation of mouse Fraction B 
3 H-DNA (prepared as described in text) with 
mouse and rat DNAs to Cot 16 in 0.14 M PB 
at 50° C. Eighty percent of the 3 H-DNA reas- 
sociated with mouse DNA and 74% with rat 
DNA. Bound material was eluted from HAP 
in 5°C steps in 0.14 M PB, 0.02% SLS. 



experiment with rat 14 C-DNA. Twenty- 
two percent of the total DNA eluted 
from HAP above 70° following incuba- 
tion to C t 3. Instantaneous binding 
material (25%) was removed by pass- 
age through HAP after incubation to 
C t 2 X 10 -5 . Forty-seven percent of 
the resulting Fraction A was able to 
reassociate with mouse DNA after in- 
cubation to C t 1 at 50°C in 0.14 M PB. 
This material was isolated in double- 
stranded form and SI -treated as above. 
Fifty-three percent of the radioactivity 
was rendered dialyzable. The nondialyz- 
able Fraction B was reassociated with 
rat or mouse DNA to C t 5 (86% and 
77% reassociation, respectively) and the 
two thermal elution profiles determined 
(T m with rat DNA about 76°C, with 
mouse DNA about 68°C) as shown in 
Fig. 29. Here again Fraction B consti- 
tutes about 4% of total rat DNA, and 
the differentially melting subclass some- 
what less. 

As with related satellite DNAs, these 
results imply an evolutionary history 
involving more than simple random mu- 
tation since divergence from the last 
common ancestor — for in that case ho- 

50 60 70 80 90 100 

Temperature, °C 

Fig. 29. Reassociation of rat Fraction B 
"C-DNA (prepared as described in text) with 
rat and mouse DNAs to Cot 5 in 0.14 M PB at 
50° C. Eighty-six percent of the "C-DNA re- 
associated with rat DNA and 72% with mouse 
DNA. Bound material was eluted from HAP 
in 5°C steps in 0.14 M PB, 0.02% SLS. 

mologous and heterologous melting pro- 
files would be identical. One possible 
explanation is a mechanism that could 
maintain a high degree of similarity 
within a family of repeated sequences in 
each species. The last common ancestor 
would have possessed a family, also of 
limited heterogeneity, the whole of which 
is the source of the modern rat and 
mouse families. In each of the diverging 
lines, members of the family have effec- 
tively evolved together. Alternatively, 
the present rat and mouse families might 
be amplifications of only two or more 
individual (and nonidentical) members 
of a family possessed by the last com- 
mon ancestor. 


Cooke, H. J., Evolution of the long range 
structure of satellite DNAs in the 
genus Apodemus, J. Mol. Biol., 94, 87, 

Gall, J. G., and D. D. Atherton, Satellite 
DNA sequences in Drosophila virilis, 
J. Mol. Biol., 85, 633, 1974. 

Hutton, J. R., and J. G. Wetmur, Length 
dependence of the kinetic complexity 
of mouse satellite DNA, Biochem. 
Biophys. Res. Comm., 52, 1148, 1973. 

Rice, N. R., Change in repeated DNA in 
evolution, in Brookhaven Symposia in 
Biology, No. 23, H. H. Smith, ed., 
Gordon and Breach, New York, p. 44, 

Rice, N. R., and N. A. Straus, Related- 
ness of mouse satellite DNA to DNA 
of various Mus species, Proc. Nat. 
Acad. Sci. USA, 70, 3546, 1973. 

Stavnezer, E., Ph.D. thesis, Johns Hop- 
kins University, 1971. 

Sutton, W. D., and M. McCallum, Re- 
lated satellite DNAs in the genus 
Mus, J. Mol. Biol., 71, 633, 1972. 

Are There Mouse Satellite Sequences 
in Main-Band RNA? 

Nancy R. Rice 

Mouse satellite DNA, which is com- 
posed of several million very similar 



copies of a sequence of 50-100 nucleo- 
tides (Waring and Britten, 1966; Rice, 
1974), is known to be localized around 
the centromeres of the mouse chromo- 
somes (Jones, 1970; Pardue and Gall, 
1970). Satellite sequences exist in a 
tightly clustered arrangement. From 
studies of the effect of molecular weight 
on percent satellite observable in CsCl 
density gradients, it was concluded that 
other, interspersed sequences can account 
for no more than about 10% (and pos- 
sibly much less) of total isolated satellite 
DNA (Rice, 1974). How did such a 
highly nonrandom arrangement arise? 
Did mouse satellite originate in one or 
more noncentromeric regions and even- 
tually, through rapid translocation, be- 
come concentrated at the 20 centromeres? 
Or did it originate in its present loca- 
tions, from which dispersal by translo- 
cation must be quite slow? One approach 
to this problem is to investigate in a 
number of different DNAs how the in- 
terspersion of satellite and main-band 
sequences varies with satellite age. I 
here report progress toward quantifying 
the extent of such interspersion for 
mouse satellite DNA. 

In the appropriate Ag+-Cs 2 S0 4 den- 
sity gradient, DNA molecules consisting 
solely of satellite DNA fail to bind Ag+, 
while the rest of mouse DNA binds sig- 
nificantly and hence increases in density 
(Fig. 30). Molecules that consist of 
partly satellite and partly nonsatellite 
would be expected to show intermediate 
binding and to be separable in varying 
degrees from pure satellite. Hence, one 
measure of the extent to which satellite 
sequences are interspersed is the degree 
to which they can be found linked to non- 
satellite DNA in such density gradients. 

Experimentally I have asked whether 
nonradioactive main-band DNA contains 
sequences able to reassociate with trace 
amounts of isolated mouse satellite 3 H- 
DNA. From the extent of such reassoci- 
ation at defined C Q t, the satellite content 
of the unlabeled DNA can be calculated. 
Fifty-five /xg main-band DNA from the 

gradient shown in Fig. 30, for example, 
was reassociated in C t 6 X 10~ 2 in the 
presence of satellite 3 H-DNA (0.01 /xg, 
C t 10- 5 , self-reaction, 22%), and 64% 
of the 3 H cpm were found to bind to 
hydroxyapatite. From the simple second- 
order reassociation curve of isolated 
satellite presented in last year's report 
(instantaneous binding, 7% ; final extent 
of reassociation, 95%), it follows that 
there is about eightfold higher satellite 
content in the incubation mix than in 
the satellite 3 H-DNA alone and that 
therefore the main-band DNA contains 
7 X -01 = .07 fig satellite, or .07/55 = 
.13% satellite. All main-band DNAs 
from several other gradients yielded 
similar results, incubation to C t 10 -2 

260 m/z 


Fraction number 

Fig. 30. 2.3 mg mouse DNA in 0.01 M 
sodium tetraborate, pH 8.9, was incubated with 
AgN0 3 at Ag + /DNA-P = 0.3 for several hours 
at room temperature. CS2SO4 was added to give 
a refractive index of 1.3722, and the solution 
was divided among four 40 rotor tubes and 
centrifuged at 33,000 rpm for 3 days at 19°C. 
Fractions were collected following tube punc- 
ture, and O.D .260 mji measured after dilution 
with 0.01 M Tris, pH 8.8. (solid circles) Main 
band DNA (fractions 7-23 in the above gradi- 
ent) from all four tubes was pooled and dia- 
lyzed against 0.01 M Tris, 0.01 M NaCl, 0.001 
M EDTA (TNE). Crosses represent density 
(g/cm" 3 ). 



to 6 X 10 -2 m the presence of satellite 
3 H-DNA resulting in 37-79% of 3 H cpm 
binding to HAP. Thus, satellite com- 
prises 0.1% to about 0.7% of these main- 
band DNAs, amounting to up to 10% 
of total cellular satellite content. 

Since this apparent interspersion may 
result from technical problems of cen- 
trifuging and fractionating, I have asked 
to what extent these satellite sequences 
remain in main-band DNA through a 
second cycle of centrifugation. Main- 
band DNA of Fig. 30, for example, was 
recentrifuged in a CsCl density gradient, 
and fractions across the gradient were 
assayed for satellite content as above. 
As shown in Fig. 31, the distribution of 
satellite is clearly different from that of 
main-band DNA. Satellite sequences are 
of lower than average density, about 
80% of the total occurring in the less 
dense half of the DNA, and about 50% 
in the least dense quarter. Since pure 
satellite exhibits very similar banding 
behavior in such gradients, it is clear 
that the initial estimate of interspersion 
is much too high. 

Recentrifuging main-band DNA from 
Fig. 30 in a second Ag+-Cs 2 S0 4 gradi- 
ent is even more decisive. As shown in 
Fig. 32 the majority of satellite se- 
quences bind little or no Ag+ and appear 
quite apart from the main-band DNA. 
Since these molecules traveled with 
main-band DNA in the first gradient, it 
appears that most of the apparent inter- 
spersion results from technical limita- 
tions. Satellite content of this second 
main band is about six times lower than 
that of the first, and thus amounts to 
about 0.3% of total cellular satellite 
DNA. I would predict that a third cen- 
trifugation cycle will reduce this even 
further. For the present, however, taking 
0.3% as the upper limit for satellite in- 
terspersion, it can be calculated that up 
to about 15,000 of the more than four 
million satellite sequences could be linked 
to nonsatellite molecules. 


Bond, H. E., W. G. Flamm, H. E. Burr, 
and S. B. Bond, Mouse satellite DNA, 
further studies on its biological and 


20 30 40 

Fraction number 


Fig. 31. 300 fig, main-band DNA from Fig. 30 was dialyzed against TNE and then centrifuged 
in a preparative CsCl density gradient for 3 days at 33,000 rpm at 19°C in the 40 rotor. Frac- 
tions were collected following tube puncture. O.D. 2 eo m^ of diluted fractions was determined, 
and indicated fractions were pooled and dialyzed against 0.01 M phosphate buffer (PB). Each 
pool was reassociated with 0.01 /xg mouse satellite 3 H-DNA (self-reaction 22%, Cot 10" 5 ) in 
0.14 M PB at 50° C, and the extent of binding of 3 H cpm to hydroxyapatite determined. 
Satellite content of each pool was calculated as described in the text. 



260 m/i 



Fraction number 

Fig. 32. 300 /xg main-band DNA from Fig. 30 was dialyzed against TNE, then against several 
changes of 0.002 M sodium tetraborate. The solution was brought up to 0.01 M sodium tetra- 
borate, incubated with AgNOs at Ag + /DNA-P = 0.3 for several hours at room temperature, 
then centrifuged with Cs 2 S0 4 at 33,000 rpm for 3 days at 19° C. Fractions were collected follow- 
ing tube puncture. Diluted fractions of O.D. 26 o mix were pooled, dialyzed against TNE, and 
assayed for satellite content as described in Fig. 31. 

physical characteristics and its intra- 
cellular location, J. Mol. Biol, 27, 289, 

Jones, K. W., Chromosomal and nuclear 
location of mouse satellite in indi- 
vidual cells, Nature, 225, 912, 1970. 

Pardue, M. L., and J. G. Gall, Chromo- 
somal localization of mouse satellite 
DNA, Science, 168, 1356, 1970. 

Rice, N. R., Fine structure in mouse 
satellite DNA, Carnegie Inst. Wash. 
Year Book 73, p. 1088, 1974. 

Waring, M., and R. J. Britten, Nucleo- 
tide sequence repetition: a rapidly re- 
associating fraction of mouse DNA, 
Science, 154, 791, 1966. 

Repeated DNA and Speciation 

R. B. Roberts, N. R. Rice, 
and N. W van de Velde 

Year Book 64 reports the discovery of 
repeated DNA, and the same Report in- 
cludes consideration of the role of re- 
peated DNA in evolution. Year Book 65 
points out that the formation of a new 
family is a more drastic and possibly 

more significant event than the slow 
changes produced by accumulation of 
mutations. The suggestion is made that 
family formation may be related to sud- 
den evolutionary changes such as bursts 
of speciation. This theme is elaborated 
in Year Book 67 which quotes G. C. 
Simpson who indicates that sudden dras- 
tic changes are evident in the fossil 

This year a report by S. M. Stanley 
provided the clue for a partial test of 
the hypothesis that family formation is 
necessary (but perhaps not sufficient) 
for speciation. He has calculated that 
the rate of speciation is three times more 
rapid in mammals than in molluscs. Ac- 
cordingly, if family formation is neces- 
sary for speciation, we would expect to 
find three times more repeated DNA in 
mammals than in molluscs. 

Accordingly, we have extracted DNA 
from clams and oysters to determine 
whether this expectation is borne out. 
Preliminary measurements show rela- 
tively little repeated DNA in these 
organisms, 14% in the oyster and 19% 



in the clam (compared to 30%-50% in 
most mammals). Furthermore, these low 
figures may be too high if the DNA used 
is contaminated by DNA from bacteria 
or other microorganisms growing within 
the molluscs. 

Thus, the early results are consistent 
with the hypothesis that family forma- 
tion is necessary for speciation. Mea- 
surements of many other organisms 
would be needed to establish a firm 
correlation between the rate of speciation 
and the rate of family formation. Ordi- 
narily we would not include such pre- 
liminary observations in our report. 
However, as this may be the last report 
of the Biophysics Section, they are in- 
cluded in the hope that someone may be 
stimulated to carry out the needed work. 


Stanley, S. M., A theory of evolution 
above the species level, Proc. Nat. 
Acad. Sci. USA, 72, 646, 1975. 

A Distinctive Feature of the 
Human Genome 

Bill H. Hoyer, Richard B. Roberts, Nancy R. 
Rice, and Neltje W . van de Velde 

Other evidence is consistent with the 
hypothesis that the formation of new 
families may be necessary for the forma- 
tion of new species. In several closely 
related rodents, families of DNA se- 
quences are found which react to form 
highly stable complexes with homologous 
DNA but which react to give only quite 
unstable complexes with the DNA of the 
near relative. This situation has been 
described in Year Book 71, p. 264; 72, 
p. 197; and 73, p. 1098. This year we 
have examined the human DNA to de- 
termine whether it contained any such 
DNA different from the DNA of the 
closely related chimpanzee. 

Highly radioactive human DNA was 
incubated to C t 100 and then adsorbed 
on hydroxyapatite and eluted in 2.5° 
temperature steps. The material which 
was highly stable (eluting at 87.5°) was 

collected and any instantaneous binding 
material removed by passing through the 
column after an extremely short incuba- 
tion. The resulting fraction was then 
reacted with an excess of nonradioactive 
DNA from human, chimpanzee, or E. 
colt. After incubation to C t 100 the three 
different samples were adsorbed to the 
column and eluted by a temperature 
gradient. The elution diagrams showed 
two major peaks, one at 72° and one at 
85°. Both are present after incubation 
with E. coli DNA and can be attributed 
to self-reaction among the tracer mole- 
cules. Incubation with chimpanzee DNA 
increases only the 72° peak, while in- 
cubation with human DNA results in 
a large increase in the 85° peak. We 
believe that this fraction is due to a 
new family of DNA sequences which was 
formed after (or possibly as the cause 
of) the divergence of the human and 
chimpanzee lines. 

This experiment has been repeated in 
slightly different circumstances three 
times but much remains to be done. 
Ordinarily such preliminary observation 
would not be reported at this time. How- 
ever, it is of considerable interest to 
find a fraction of the human DNA (ap- 
proximately 1%) that does differ mark- 
edly from that of chimp. The bulk of 
human DNA differs from chimpanzee 
in only 1% of the nucleotides. Since 
these findings may not be followed up 
in this laboratory, we considered that 
the results should be reported. 

Protein-Protein Interactions 

Dean B. Cowie, Georges N. Cohen* 
and Paolo Truffa-Bachi* 

Cellular activities are regulated by 
two main mechanisms. One, operating 
at the genetic level, controls the rate of 
translation of genetic material into spe- 
cific messenger RNAs and has been the 
subject of many studies which can be 

* Institut Pasteur, Service de Biochimie 
Cellulaire, Paris, France. 



grouped under the generic name induc- 
tion-repression mechanisms. The second 
regulates the activity of already synthe- 
sized proteins by modifying them either 
covalently (phosphorylation, adenyla- 

changes in the environment of specific 
aromatic residues, by difference spectros- 
copy or fluorescence. Many of these 
methods have been applied to the en- 
zyme under study. We have focused our 

tion, ADP-ribosylation, etc.) or by attention on the conformational changes 
changing their conformation by an alio- brought about by the reaction of the 
steric transition under the influence of protein with specific antibodies, and on 
small ligands which act as modulators or qualitative and quantitative effects of 
transducers. These covalent or confor- the various ligands of the protein on this 
mational changes lead to an increase or 
decrease of catalytic activity, always in 
the "interest" of the cell economy. We 
have been interested in the latter mech- 
anism for several years and have chosen 


Activity of the Antibody -bound Enzyme 
Antibody binding inhibits the activity 

to study the conformational changes of some enzymes completely, some par- 
caused by ligand binding (the ligand tially, and others not at all. In some 

being either substrates or feedback in- 
hibitor) to aspartokinase I-homoserine 
dehydrogenase I of Escherichia coli, an 
enzyme occupying a key position in the 
biosynthesis of several essential metabo- 

The ways to detect changes are nu- 
merous, from the ones that detect gross 
conformational changes, such as optical 

instances, antibody-induced enzyme ac- 
tivation has been reported (Cinander, 
1967; Cowie et al, 1961). 

We have investigated the interaction 
of the allosteric enzyme aspartokinase 
I-homoserine dehydrogenase I (AKI- 
HDHI) with antibodies specific for this 
protein. This report concerns some of 
the characteristics of this antibody- 

rotatory dispersion or hydrogen-deute- antigen complex in terms of the activity 
rium exchange, to those which detect of the homoserine dehydrogenase and 

TABLE 3. Reduction of Homoserine Dehydrogenase Activity upon Binding of Aspartokinase 
I-Homoserine Dehydrogenase I to its Specific Immunoadsorbent* 

Experiment Number 

Units Bound Enzyme 

Units Enzyme Activity 
on the Column (maximum) 





























av. = 28.8 

* Various nonsaturating quantities of AKI-HDHI were loaded on the immunoadsorbent 
column in buffer (0.5 ml-2.0 ml, pH 7.2). Buffer washes (70-300 ml) were then percolated 
through the column, and the HDH activity of the eluate fractions assayed. The total activity 
found in the eluates is subtracted from the units loaded on the column to determine the units 
of antibody-bound enzyme. The quantity of eluted enzyme is usually only a few percent of the 
quantity loaded. The column volume was 2.0 ml, and the diameter of the column was 2 cm. 
HDH activity of the antibody-bound enzyme was determined by percolating substrate solu- 
tion through the column at a fixed rate (see text) . 



its threonine sensitivity. HDH activity 
is measured in the direction HS -> ASA 
by the method of Patte et al., using 
HDH substrate solution: 10 ml M Tris, 
pH 9.0; 25 ml M KC1; 10 ml 2.5 M L- 
homoserine; and 10 mg NADP. 

Immunoadsorbent columns were pre- 
pared containing antibodies specific for 
AKI-HDHI (Cowie et al, 1973). Usu- 
ally a column was loaded with a non- 
saturating quantity of purified enzyme 
and washed with buffer (20 ml potas- 
sium phosphate, pH 7.2, and 0.15 M 
KC1) to remove unbound enzyme. The 
homoserine dehydrogenase activity and 
the threonine sensitivity of the antibody- 
bound enzyme can be easily determined 
by flowing substrate solution (in the 
presence or absence of feedback inhibi- 
tor) through the column and measuring, 
immediately after elution, the absorbency 
of the eluates at 340 nm. The enzyme 
activity on the column can be calculated 
from the absorbency values and the flow 
rate of the substrate solution (Feldmann 
et al, 1972; Cowie et al, 1974). From 
the kinetic studies presented below it 
can be seen that this immunoadsorbent 
method has unique advantages for the 
study of slow kinetic reactions. 

Table 3 demonstrates that only a por- 
tion of activity of the homoserine de- 
hydrogenase loaded on the column re- 
mains active. Figure 33 shows that a 
maximum of enzyme activity is observed 
soon after the substrate solution is added 
to the column. In some instances, where 
the initial observed enzyme activity ex- 
ceeds 25% of the load (Table 3), there 
is a slow decrease in the column activity 
until a value is reached equal to approxi- 
mately 25% of the activity of the load 
(Fig. 33). The activity of the antibody- 
bound enzyme does not appear to be 
sensitive to enzyme concentration (over 
the range studied, Table 3). 

The antibody-bound enzyme is only 
slightly sensitive to L-threonine inhibi- 
tion. Figure 33 shows that there is ap- 
proximately a 10% inhibition when the 
substrate solution is supplemented with 







































rwv_7 ' 






















— 1 

' l_ 







1 Substrate 


P / after 
¥ 1 

buffer washes 

1 1 1 

20 40 60 

Time in minutes 



Fig. 33. Homoserine dehydrogenase activity 
observed on the antibody column during per- 
colation with substrate solution. After loading 
with a non-saturating quantity of HDH, the 
column was washed for 70 minutes (300 ml 
buffer, pH 7.2) before the percolation with 
substrate solution was started. After 40 min- 
utes the substrate solution was supplemented 
with L-threonine (2 X 10" 3 M final concentra- 
tion) ; at 60 minutes unsupplemented substrate 
solution was employed. The flow rate of the 
substrate solution was 0.9 ml/min, and samples 
were taken every 5 minutes. At the time the 
substrate solution was first added to the anti- 
body column the quantity of bound HDH was 
2200 units; the percentages given are the ob- 
served HDH activities on the column relative 
to this quantity of bound enzyme. Other col- 
umn details are as given in Table 3. 

L-threonine (final concentration 2 X 
10 -3 M). Restoration of the antibody- 
bound enzyme activity (—10%) occurs 
when the threonine is removed. 

During the passage of substrate 
through the column there is little re- 
moval of active homoserine dehydro- 
genase from the antibody-enzyme com- 
plex. This can be shown by measuring 
the change in absorbency, with time, of 
each eluate fraction after the initial 



absorbency determination has been 
made. Only a few percent of the activity 
of the enzyme load appears in the eluate 
washes, and for the most part in the 
early eluate fractions. 

It is evident that the antibody-bound 
enzyme exists in a conformational state 
quite different from that of the free 
enzyme in buffer — where the threonine 
inhibition is 70% and the enzyme, in the 
absence of threonine, is fully active. 

When the same type of kinetic experi- 
ments are repeated with the substrate 
solution supplemented with 2 X 10~ 3 
M L-threonine, completely different re- 
sults are obtained. 

Figure 34 shows that under these con- 
ditions the homoserine hydrogenase ac- 
tivity of the antibody-bound enzyme can 

initially be as much as 50% of the 
activity of the free enzyme. Further- 
more, removal of the threonine from the 
percolating substrate solution results in 
a slow inactivation of the antibody- 
enzyme complex, which then can be 
partially reactivated by addition of 
threonine to the substrate solution. 

These results suggest that the enzyme, 
loaded and bound to the antibody (in 
buffer at pH 7.2) can slowly assume a 
new conformation state which in pres- 
ence of threonine and the other ligands 
contained in the substrate solution (at 
pH 9.0) is more active than the free 
enzyme in the presence of the same 
amount of threonine. That threonine 
plays a major role in such a conforma- 
tional transition can be demonstrated 

~ 50- 

R 40 


* 20 

• = Substrate + Threonine 
o = Substrate alone 



Time in minutes 


Fig. 34. Homoserine dehydrogenase activity observed on the antibody column following 
percolation with threonine supplemented (2 X 10" 3 M) substrate solution. Buffer wash pro- 
cedure as in Fig. 33. At the times indicated, unsupplemented (or threonine-supplemented) 
substrate solutions were used for further percolation. A continuous flow rate of 2.7 ml/minute 
was used, and samples were taken every 2 minutes. The quantity of bound HDH at the end 
of the buffer wash was 1400 units. The percentages given in this figure are the observed HDH 
activities relative to this quantity of bound enzyme. Other column details are as given in 
Table 3. 



by the data of Fig. 34 in which only the 
removal of threonine from the substrate 
solution causes a reduced activity of the 
antibody-enzyme complex. 

A more dramatic presentation of the 
phenomenon is shown in Fig. 35. Here, 
repeated oscillations in enzyme activity 
are produced by the alternative removal 
and restoration of threonine to the sub- 
strate solution. 

It should be noted that the effect of 
the threonine supplementation diminishes 
with time. In addition, the time required 
to go from the most to the least active 
conformational state is of the order of 
10 minutes. In these experiments, the 
flow rate of the percolating substrate 
solution was 2.3 ml per minute; the vol- 
ume of the column was 2.0 ml; and each 
sample took 2 minutes to collect. Thus, 
in less than a minute all of the indi- 
vidual enzyme-antibody complexes were 
exposed to each change of environmental 

During the course of this experiment 

there was some loss of active enzyme to 
the eluates. The total active enzyme 
lost during the 90 minutes amounted to 
10% of the quantity of enzyme initially 
bound to the antibody column. 

The results reported above are mark- 
edly different from those reported last 
year (Cowie et al., 1974) when similar 
studies were carried out. The only ex- 
perimental difference between the two 
sets of experiments is the buffer wash 
procedure which was added to this year's 
studies. Very little enzyme is eluted by 
the buffer wash, and it has been con- 
cluded that the time the enzyme remains 
bound to the antibodies (before assaying 
with substrate solution) has a significant 
bearing on the state of the enzyme. This 
conclusion is supported by the following 

The minimal time consumed by the 
buffer wash procedure was 70 minutes. 
When a nonsaturating load of enzyme 
is added to the column in buffer and 
allowed to stand for 70 minutes before 





S 30 



o 1 * 

• = Substrate + Threonine 
o = Substrate alone 





Time in minutes 

Fig. 35. Homoserine dehydrogenase activity observed on the antibody column following 
percolation with threonine-supplemented (2 X 10~ 3 M) substrate solution. Buffer wash pro- 
cedure as in Fig. 33. At times indicated, supplemented or threonine-supplemented substrate 
solutions were used for percolation. A continuous flow rate of 2.3 ml/minute was used, and 
samples were taken every 2 minutes. The quantity of bound HDH at the end of the buffer 
wash was 1400 units. The percentages given in this figure are the observed HDH activities 
relative to this quantity of bound enzyme. Other column details are as given in Table 3. 



assaying with substrate solution, the 
characteristics of the antibody-bound 
enzyme are found to be identical to those 
shown in Fig. 33. The HDH activity was 
about 25% of the activity of the enzyme 
load and only slightly sensitive to 

Additional studies, investigating the 
state of the antibody-bound enzyme 
immediately after the enzyme has been 
loaded on the column, gave results typi- 
cal of those shown in Fig. 36. 

A maximum of enzyme activity is 
observed soon after the flow of substrate 
solution is initiated. This is followed by 
a rapid loss of antibody-bound enzyme 
activity which eventually decreases 
asymptotically to a slower rate. The max- 
imal enzyme activity observed on the 
column varies between 55% and 70% 
and depends upon the flow rate of sub- 
strate solution. The optical density of 
the eluate increased with time, showing 
that enzyme is released from the column 
(Fig. 36 inset). 

The biphasic nature of the data (Fig. 
36) indicates that the loss of activity 
of the antibody-bound enzyme involves 
more than a single process. The appear- 
ance of active enzyme in the eluates in- 
dicates that some dissociation of the 
enzyme from antibody binding does 
occur when substrate solution (at pH 
9.0) is percolated through the column. 
However, sufficient active enzyme is not 
recovered in the eluate fractions to ac- 
count for the loss of activity of the anti- 
body-enzyme complex. During the first 
30 minutes of this experiment the 
observed loss of the column activity 
(maximum value, 56% of load activity) 
dropped by 31% of the activity loaded 
on the column. Over this period of per- 
colation of substrate, only 15% of the 
total load of enzyme activity was re- 
covered in the eluate fractions. 

Initially the antibody-bound enzyme 
is sensitive to threonine; however, dur- 
ing the inactivation process, there is a 
corresponding decrease in threonine sen- 
sitivity. At the end of 50 minutes the 



Tirne in minutes 

Fig. 36. Homoserine dehydrogenase activity 
observed on the antibody column following 
percolation with substrate solution which was 
added immediately after loading the column 
with HDH. The buffer wash procedure was 
eliminated. The flow rate was 2.7 ml/minute, 
and samples were taken every minute. The 
inset figure shows the cumulative values of 
actual enzyme appearing in the eluate during 
the course of the experiment. The percentages 
given in this figure are the observed HDH 
activities relative to the activity of the load 
of 2800 units of HDH. Other column details 
are as given in Table 3. 

threonine inhibition is only about 10%. 

The above results suggest that as soon 
as the antibody-enzyme complex is 
formed, inactivation of HDH is initiated. 
Furthermore, the substrate solution (pH 
9.0) dissociates a portion of the enzyme 
from the antibodies, a process not ob- 
served when buffer (pH 7.2) is percolated 
through the column. The activity of the 
enzyme-antibody complex, even in buf- 
fer at pH 7.2, however, decreases from 
the time the enzyme becomes bound to 
the antibody column. 

Figure 37 shows the results of another 
experiment in which substrate solution, 
supplemented with 2 X 10~ 3 M L- 
threonine, is percolated through the col- 







Substrate alone 


Substrate + Threonine 


20 25 

30 35 40 45 

Time in minutes 

Fig. 37. Homoserine dehydrogenase activity observed on the antibody column following 
percolation with threonine-supplemented (2 X 10 3 M) substrate solution. The buffer wash 
procedure was eliminated. The flow rate was 2.7 ml/minute, and samples were taken every 
minute. At 36 minutes unsupplemented substrate solution was used. The percentages given in 
this figure are the observed HDH activities relative to the activity of the load of 900 units of 
HDH. Other column details are as given in Table 3. 

umn immediately after the addition of a 
nonsaturating enzyme load. Under these 
conditions, the loss of activity of the 
antibody-enzyme complex can be repre- 
sented as a single exponential. Further- 
more, removal of the threonine from the 
substrate results in an increase of HDH 
activity (Fig. 37). This increase reaches 
a value which, when compared to the 
activity on the column immediately be- 
fore the removal of threonine, shows that 
the activity remaining on the column is 
normally inhibitable by threonine. The 
same degree of recovery of enzyme ac- 
tivity with the removal of threonine 
from the substrate solution occurs over 
the entire course of the experiment. Thus, 
other than the overall loss of enzyme 
activity on the column, the antibody- 
enzyme complex reacts to threonine just 
like the soluble enzyme. 

As in the previous experiment, Fig. 36, 
the recovery of active enzyme in the 
eluate fractions is inadequate to com- 
pensate for the observed loss of column 

HDH activity. About 15% of the ac- 
tivity was eluted during the first 36 
minutes of substrate percolation; the 
decrease in column activity over the 
same period of time was 33%. This in- 
dicates that either inactive protein is 
being eluted or that the antibody-bound 
enzyme is being inactivated during the 
time of the experiment (or that both 
processes are occurring) . Since the quan- 
tity of protein eluted in these experi- 
ments is so low, its determination de- 
pends upon the use of labeled enzyme 
preparations. Such a study is now under 

One can conclude from these studies 
that there are conditions under which 
the antibody-enzyme complex can be- 
have like the free enzyme in solution. 
This conformation is observed before 
the enzyme has had sufficient time to 
undergo a conformational change im- 
posed upon it by its antibodies and is 
observable following loading of the 
enzyme on the antibody column and 



percolating immediately with substrate 
solution (db L-threonine). 

Conversely, if the enzyme is allowed 
to react for a sufficient time with the 
antibodies, it undergoes gradually an 
irreversible conformational change, re- 
sulting eventually in the loss of 75% to 
80% of the original activity and in the 
loss of its sensitivity to threonine. Dur- 
ing the process of the temporal inactiva- 
tion by the antibodies the addition of 
threonine retards the ultimate formation 
of the irreversible configuration, pro- 
ducing a paradoxical increase in the 
apparent conformation of the enzyme 
less susceptible to antibody inactivation. 

The use of antibodies has allowed us 
to probe some of the conformational 
changes of the enzyme other than those 
imposed upon it by the allosteric equi- 
libriums which can be displaced by its 
specific ligands. It is also felt that such 
studies may provide some information 
regarding the antigen- antibody reaction 


Cinander, B., Antibodies to enzymes — 
a discussion of the mechanisms of in- 
hibition and activation, Proc. 2nd 
Meeting of the Federation of European 
Biochemical Societies, p. 85, Vienna, 
Austria, April 21-24, 1965. 

Cowie, D. B., S. Spiegelman, R. B. Rob- 
erts, and J. D. Duerksen, Ribosome- 
bound-galactosidase, Proc. Nat. Acad. 
Sci. USA, 47, 114-122, 1961. 

Cowie, D. B., P. Truffa-Bachi, J. M. 
Costrejean, M. C. Py, and G. N. 
Cohen, A preliminary immuno-chemi- 
cal study of E. coli aspartokinase I- 
homoserine dehydrogenase I. Biochem. 
Biophys. Res. Comm., 53, 188-193, 

Cowie, D. B., G. N. Cohen, and P. 
Truffa-Bachi, Protein-protein inter- 
action, in Carnegie Inst. Wash. Year 
Book 73, pp. 1064-1088, 1974. 

Feldmann, K., H. Zeisel, and E. Helm- 
reich, Interaction between native and 
chemically modified subunits of ma- 

trix-bound glycogen phosphorylase, 
Proc. Nat. Acad. Sci. USA, 69, 2278- 
2282, 1972. 

Historical Review of 
Biophysics Section 

R. B. Roberts 

Formation of the Section 

At this time, when the 30-year exist- 
ence of the Biophysics Section is in the 
final stage of termination, it seems ap- 
propriate to look back at its history and 
accomplishments. A Biophysics Section 
was first authorized by Vannevar Bush 
in 1946 during M. A. Tuve's first year 
as Director of the Department. Its ori- 
gins can be traced further back. The 
first publication in the area of biophysics 
appeared in 1931 on "Biological Effects 
of Gamma Rays." This subject was 
timely in anticipation of the exposure of 
laboratory workers to intense sources of 
radiation which would be produced when 
the disintegration of nuclei by artificially 
accelerated particles was achieved (1932) . 

Following this discovery and the dis- 
covery of artificial radioactivity (1934), 
nuclear physics equipment and tech- 
niques became very attractive to the 
biological community. During the late 
1930's the Department had available 
two of the relatively scarce accelerators, 
and there is no wonder that biologists 
wanted to share these facilities. Effects 
of gamma rays, neutrons, and charged 
particles on all sorts of biological ma- 
terial needed study. More importantly, 
radioactive isotopes of most elements 
could be produced and used as tracers 
to answer questions previously unap- 
proachable. Many such experiments were 
carried out using the two Van de Graaff 
accelerators then in operation. As exam- 
ples, fruit flies were bombarded with 
neutrons for Demerec ; in 1938 24 Na was 
produced and used to study the flow of 
material from mother to embryo in rats 
(this study is now reprinted in Classic 
Papers in Obstetrics and Gynecology) ; 


n C was used to study the uptake of bacterial viruses, but following one staff 

C0 2 in photosynthesis in 1940. member's participation in the bacterio- 

In fact, these activities infringed seri- phage course at Cold Spring Harbor 
ously on the time available for nuclear there was a steadily increasing use of 
physics. Tuve then persuaded Bush to the bacterium Escherichia coli as the 
provide funds for the construction of a best suited to our interests (see CIW 
large cyclotron with capacity to provide Publ. 624, pp. 656-690, 692-695). 
tracers for the entire Carnegie Institu- 
tion and the Washington area. The de- Biosynthesis of Small Molecules 
sign of this machine and the associated 

laboratory was started in early 1939. By 1952 our interests had focused on 

The war years then intervened, but the the synthesis of amino acids and nucleo- 

cyclotron was completed in 1942. It was tides by E. coli, and this subject was 

used during the war to produce radio- studied intensively through 1955, cul- 

active isotopes of arsenic and antimony minating in the publication of Studies of 

for studies of tropical medicine for the Biosynthesis in E. Coli. During this 

Navy, and later to produce isotopes for period the method of isotopic competi- 

general distribution. tion was introduced and used extensively. 

At the conclusion of the war Tuve As an example of this method, bacteria 
became Director of the Department and are provided with radioactive glucose as 
brought about a reorientation of the a carbon source, and an appropriate 
Department's program. In particular, nonradioactive amino acid is added to 
nuclear physics had received a tremen- the medium as a competitor. After a 
dous emphasis from the development of period of growth the cells are harvested; 
the atom bomb and seemed destined to their proteins are isolated, and the amino 
be carried out by large groups with un- acids of the proteins examined. Radio- 
limited money. In contrast, biophysics activity does not appear in the amino 
appeared to be more suitable for small acid that was added or in the other 
groups and (with the advent of tracers) amino acids derived from it. From such 
due for a burst of progress. In addition, results it is possible to deduce the path- 
our prewar experience had demonstrated ways of synthesis in the cell and to cal- 
that physicists could contribute to biol- culate the flow along each pathway, 
ogy not only in technique but in the Such calculations demonstrated the im- 
planning of experiments. Tuve therefore portance of the Kreb's Cycle (previously 
decided to reduce the emphasis on nu- recognized as a component of carbohy- 
clear physics and to initiate a Biophysics drate catabolism) in the synthetic ac- 
Section. At this initial stage the group tivities. 

included P. H. Abelson (chairman), D. Furthermore, the replacement of one 

B. Cowie, and Roberts, all trained in carbon source by another was not simply 

physics. In addition, we were fortunate due to dilution but to the abrupt termi- 

to have with us H. H. Darby and W. R. nation of synthetic activities no longer 

Duryee as guest investigators who pro- needed. These changes were described at 

vided a much needed wealth of biological the time as a chemical response and 

experience. contrasted to the biological response of 

The experiments of the first years a reduction in the synthesis of enzymes 

(1947-1951) were largely exploratory, that were no longer needed. These two 

including such diverse topics as capillary processes are now termed "feedback in- 

permeability and the effects of ultraviolet hibition" and "enzyme repression." The 

radiation, vitamin B 12 , and various ions results of these years are described in 

on biological material. The biological CIW Publication 607, Studies of Bio- 

materials used ranged from humans to synthesis in Escherichia coli (1955). 



Metabolic Pools 

Having a quite complete flow pattern 
for carbon in E. colt, we chose not to 
continue this work in other organisms 
but to investigate the properties of the 
"pools" in which the small building 
blocks of protein and nucleic acids are 
held in the cell prior to incorporation 
into the macromolecules. These com- 
pounds are concentrated by the cells, 
and it appeared that the mechanisms for 
their concentration might be related to 
the machinery for synthesis. This sub- 
ject was investigated in both E. coli and 
yeast. It turned out to be extremely com- 
plicated: Some amino acids competed 
with others for entry into the cell; some 
compounds were incorporated into mac- 
romolecules without equilibrating with 
pools already present in the cell. The 
results had to be interpreted in terms 
of what we called a carrier molecule. 
The term "permease" (used by the 
Pasteur group and now generally ac- 
cepted) was distasteful to us because 
at that time there was little evidence 
that an enzyme was involved. 

These studies turned out to yield little 
information concerning the mechanisms 
of macromolecular synthesis, but the 
knowledge of the properties of the pools 
and the kinetic delays that they intro- 
duce was essential to the proper interpre- 
tation of later studies of protein and 
ribosome synthesis. They are included 
in the material of CIW Publication 624, 
Studies in Macromolecular Biosynthesis, 

Experimental work on the properties 
of the pool did not terminate abruptly; 
in fact, further information became 
essential and our last paper on pools was 
not published until 1963. After 1957, 
however, interest shifted increasingly to 
larger components of the cell. At that 
time the literature carried reports that 
a fraction had been isolated from dis- 
rupted animal cells which contained 
ribonucleoprotein particles mixed with 
fragments of membranes. This fraction 
had also been shown to be rich in newly 

formed protein. Similar ribonucleopro- 
tein particles (now termed ribosomes) 
had been found in bacteria. Our atten- 
tion shifted to these particles and their 
possible involvement in protein synthesis. 


By 1959 methods had been developed 
to keep the particles stable for frac- 
tionation either by sedimentation or on 
DEAE columns. However, we found no 
indication that the particles carried 
newly formed protein. The difficulty lay 
in the small quantity and rapid turn- 
over of the nascent protein. Further ex- 
periments were carried out (with the 
participation of K. McQuillen) in which 
the period allowed for protein synthesis 
was decreased from minutes to a few 
seconds and the nascent protein was 
clearly shown in association with the 
ribosomes. This experiment showed for 
the first time that the machinery of pro- 
tein synthesis consisted of ribosomes and 
not the other components of the micro- 
some fraction. Further work along these 
lines (with S. Spiegelman) showed a 
small fraction of beta-galactosidase as- 
sociated with ribosomes. Furthermore, 
these ribosomes could be precipitated by 
addition of antibody to the enzyme fol- 
lowed by addition of antibody to the first 
antibody. Thus a method was available 
to select a particular class of ribosome. 
Currently this procedure is used in 
radioimmunoassays and RNA messenger 

Numerous other studies of ribosomes 
culminated in a series of four papers 
which described the kinetics of syn- 
thesis of ribosomes including both the 
nucleic acid and the protein components. 
Two sequential classes of precursors 
(which we called eosomes and neosomes) 
were clearly observable and seemed to 
account for most of the newly formed 
RNA of the cell. 

Just at that time the messenger theory 
of Monod was promulgated. This theory 
predicted that newly formed RNA had 
only a transient existence and was 



broken down after serving as template 
for protein synthesis. The base composi- 
tion of the newly formed RNA was 
predicted to be similar to DNA and 
different from ribosomal RNA and RNA 
of DNA-like composition that had been 
observed under special conditions. This 
did not agree with our observations that 
most newly formed RNA was ribosomal 
precursor. This discrepancy was resolved 
by a series of experiments (Midgley) 
that measured the composition of RNA 
in a series of bacteria of different DNA 
compositions. The RNA changed in 
composition with time of exposure to the 
tracer, indicating that roughly one-third 
was DNA-like RNA and had a rapid 
turnover. The remaining two-thirds of 
the RNA was stable ribosome precursor. 
These findings should have settled the 
issue; one-third of the newly formed 
RNA had the properties predicted for 
messenger. However, for a number of 
years thereafter, newly formed RNA 
was considered to be entirely messenger 
by most workers in the field. 

More direct evidence was needed to 
convince the doubters since measure- 
ments of composition could detect the 
messenger but did nothing to isolate it. 
Bolton and McCarthy then devised a 
method for immobilizing single-stranded 
DNA by trapping it in agar. RNA could 
then be incubated with the single- 
stranded DNA, and RNA strands could 
be bound to its complementary DNA. 
This procedure allowed the separation 
of messenger RNA from ribosomal RNA 
since there were relatively few sites 
where the ribosomal RNA could be 
bound. Unbound RNA was then washed 
out and the bound RNA recovered by 
washing with a different solution or at a 
higher temperature. Messenger RNA was 
isolated in useful quantities free from 
the much more abundant ribosomal 
RNA. This method brought about a 
notable advance in the study of DNA- 
RNA interactions as the earlier methods 
were either tedious or not generally 

The development of the agar column 
brought about a major shift in the pro- 
gram of the section as it rapidly became 
apparent that here was a procedure 
which allowed exploration of entirely 
new areas. These developments will be 
described below. 

In addition to the main theme of ribo- 
somes, ribosome synthesis, and messenger 
RNA outlined above, many other aspects 
of protein synthesis were studied con- 
currently. These include work on enzyme 
induction and the incorporation of ana- 
logs. This experimental work (up to the 
use of the agar column) is gathered to- 
gether in Studies in Macromolecular Bio- 
synthesis, CIW Publication 624 (1964). 

Nucleic Acid Interactions 

The procedure for trapping DNA in 
agar was developed in the spring of 1962, 
and its use in isolating messenger RNA 
is briefly mentioned in Year Book 61. 
In the following year the method was 
used extensively. The kinetics of syn- 
thesis of messenger RNA was observed, 
and the relative proportions of different 
classes of messenger RNA were esti- 
mated. The reaction specificity was de- 
termined, and it was found that unre- 
lated RNA would not bind, but RNA 
from closely related species did bind to 
some extent. 

The first experiments used RNA to 
bind to a DNA column; subsequently 
DNA-DNA reactions were investigated. 
This technique was utilized to show 
both the degree of relatedness among 
different bacteria and the homology be- 
tween E. coli DNA and the DNA of its 
lysogenic phage lambda. 

Reports in the literature stated that 
renaturation of DNA from animals did 
not occur. This result seemed plausible 
because of the much larger genome and 
the corresponding dilution of individual 
genes. In spite of this, highly successful 
experiments were conducted using ani- 
mal DNAs. The report in Year Book 62 
includes measurements showing the de- 
gree of relatedness for various organisms, 


from E. coli to man (see Table 4). bacteria and their phages and among 

Within the first year the versatility of animals. The evolutionary significance is 

the method and its application to a wide shown in the relationship between the 

variety of problems including taxonomy degree of divergence among DNAs as 

and evolution had been demonstrated. observed by these techniques and the 

The following year saw further use time (estimated by evolutionists) since 

and development of the agar column, the divergence of the species (see Fig. 

Competition was introduced as an addi- 38). 

tional tool. Here the similarity of two The application of the technique to a 
nucleic acids is shown by their ability medical problem is also reported. In 
to react with one another as well as to cooperation with D. Axelrod, DNA corn- 
compete for a limited number of binding plementary to polyoma virus DNA was 
sites on the entrapped DNA. Thus, a observed in mouse DNA and the amount 
large excess of E. coli DNA has no effect was strikingly increased in DNA derived 
on the binding of human DNA to human from mouse tumors induced by polyoma. 
DNA trapped in the agar. However, a This approach is widely used at present 
large excess of nonradioactive human in the search for human tumor viruses. 
DNA will reduce the binding of radio- The experimental results obtained with 
active human DNA. And so will a large the agar column are reviewed in Molec- 
excess of monkey DNA (although to a ular Genetics, edited by J. H. Taylor, 
lesser extent) . The degree of competition Academic Press, New York (1967). 
indicates the degree of similarity. One 

great advantage of this system is that Repeated DNA 

radioactive DNA from each species is In 1965 the studies of nucleic acid 

not needed. interactions took on a new aspect. Ever 

Competition was used to show that the since the first use of the agar column 
DNAs from various organs of mouse for mammalian DNA, a puzzle remained 
are identical but the RNAs of these as to why it worked at all. The mam- 
organs are quite different. This finding malian genome being 500 times larger 
provided experimental support to the than E. coli DNA, its DNA should re- 
theory of development that the genes of nature 500 times slower than bacterial 
different tissues are the same, but their DNA (at the same concentration). Yet 
expressions differ. this did not happen. Perhaps some mys- 

Competition was also used to extend terious property of the agar accelerated 

the studies of relatedness among the the reaction. Renaturation was then 

TABLE 4. Binding of DNA Fragments from Animals and 
Bacteria to Denatured DNA Embedded in Agar 


32 P-DNA Bound 


to the 

Homologous DNA 

DNA in Agar 



E. coli 










• • • 




• • • 

Guinea pig 



■ • • 




• • • 







E. coli 










Man, chimp 

: ■ 

T New world monkey 


- \ Loris 


\ Armadillo 


V— Mammal 

— \ Bird 




i i i i 


100 200 300 400 500 600 

Time (m.y.) since divergence 

Fig. 38. Relationship between polynucleotide 
similarity and time of evolutionary divergence. 

tested in solution, and it proceeded as 
rapidly as in agar. Careful studies of 
the renaturation rate showed that a 
fraction of mouse DNA reacted more 
rapidly than E. coli DNA instead of 500 
times slower. 

This finding brought Britten to the 
realization that a large fraction of the 
mouse DNA must be composed of a 
large number of copies of the same (or 
very similar) sequences. In fact, calcu- 
lations indicated that the rapidly re- 
naturing fraction (about 10% of the 
total) contains about one million copies 
of a short sequence of DNA. Of the 
remaining DNA, roughly half is found 
in families of 1000 to 100 members, and 
the rest has only one or very few copies 
per genome. 

These results answered one question 
but raised a host of others. How widely 
was this repeated DNA distributed 

through the various species of organisms? 
How did it originate? What role did it 
have in evolution? A new technique was 
provided for fractionating DNA accord- 
ing to its rate of renaturation. What were 
the properties of these fractions? The 
earlier agar experiments appeared to 
have observed only the repeated DNA 
and its products. Would unique DNA 
have given the same results? Were the 
members of the families identical or just 
similar enough to react? 

Since 1965 a large share of the efforts 
of the Biophysics Section have been de- 
voted to answering these questions. Those 
amenable to direct proof have been 
answered; but others are more elusive, 
requiring an accumulation of indirect 
evidence. Thus, an early survey showed 
repeated DNA in more than 50 species 
throughout the range of phylla, and it 
seemed quite safe to conclude that re- 
peated DNA made up a large fraction of 
the genomes of all eukaryotic organisms. 
Next, the unique sequences were exam- 
ined and a portion was found to be 
transcribed. Breaking the DNA into 
fragments of different sizes showed that 
the repeated fraction was distributed as 
short sequences scattered throughout the 
genome. By 1969 a summary of proper- 
ties of repeated DNA was reported (see 
Table 5). 

Subsequent developments include: (1) 
isolation of the ribosomal genes of E. 
coli; (2) formulation of a theory for the 
role of repeated DNA in controlling the 
expression of DNA; (3) demonstration 
that messenger RNA (isolated from 
polysomes) is mainly transcribed from 
unique DNA; (4) measurement of the 
rates of divergence of unique and re- 
peated DNA in rodents and primates; 
(5) study of the relatedness of the DNAs 
among bacteriophage, blue-green algae, 
and plants; (6) measurement of the 
effect of mismatching on the kinetics of 
DNA reassociation; and (7) analysis of 
properties of the fraction of eukaryotic 
DNA that appears to reassociate in- 



TABLE 5. Observed Properties of Repeated DNA Sequences 

1. Occurrence 

2. Quantity 

3. Frequency 

4. Precision 

5. Arrangement 

6. Age 

7. Variety 

8. Expression 

9. Control of 


Observed in all tested species above the fungi. 

From 20% to 80% of the total nuclear DNA. 

From 50 to 2,000,000 related sequences per family. 

All degrees of thermal stability seen in reassociated repeated DNA. 

Scattered throughout the length of the genome. 

Several hundred million years up to very recent time. 

Patterns of frequency and precision vary widely even among vertebrates. 

RNA complementary to some repeated DNA sequences has been observed 
in every cell type examined. 

Different sets of repeated sequences are transcribed in different tissues and 
stages of development. 

The Brain 

Beginning in 1955 the Biophysics Sec- 
tion has carried on a collaborative pro- 
gram with Drs. L. B. and J. B. Flexner 
of the University of Pennsylvania in 
studies of the brain. The central theme 
of this work has been to determine the 
biochemical basis of long-term memory. 
Behavior patterns can be inherited (e.g., 
bird songs) and thus must be transmitted 
to the progeny by DNA. Furthermore, 
the DNA is expressed in the phenotype 
by the kinds and quantities of protein 
synthesized. Possibly then, some be- 
havior patterns develop independently 
but others require a stimulus from the 
environment for their maturation. How- 
ever, both could have the same basic 
mechanism. Thus we attempted to test 
the hypothesis that protein synthesis, or 
protein synthesizing systems, provided a 
mechanism for memory formation. 

The early work applied the methods 
developed for study of protein synthesis 
in bacteria to the mouse brain. It was 
observed that synthesis proceeded at a 
rapid pace in the brain and that suffi- 
cient newly formed protein was available 
to account for new connections among 
neurons. Next, it was found that puro- 
mycin (a protein-synthesis inhibitor) 
could disrupt memory even when injected 
weeks after the initial training. This 
appeared to support the hypothesis. 

Soon thereafter, however, a stronger 
inhibitor of protein synthesis (acetoxy- 
cycloheximide) had no effect on mem- 
ory (if injected a day after training). 
The hypothesis seemed to have failed. 

Later it was found that memory could 
be restored in puromycin-treated mice 
by a number of different procedures, in- 
cluding injection of water into the brain 
or the use of adrenergic stimulating 
drugs, or simply by leaving the mouse 
in the cold (which stimulates the ad- 
renergic system). Radioactive puromy- 
cin-peptides were found in the brain 
even months after the administration of 
puromycin, and the procedures which 
restored memory also reduced the level 
of puromycin-peptides in the synapto- 
some fraction to levels that were not 
memory inhibiting. Puromycin blockage 
appeared to be caused by the formation 
of puromycin-peptides, which adsorbed 
to receptor sites and blocked certain 
synapses. Presumably these are receptors 
for catecholamines as the puromycin has 
a structural resemblance to these com- 

Thus, experiments designed to demon- 
strate a role for protein synthesis in 
memory formation ended in implicating 
the catecholamines. Furthermore, other 
classes of compounds are involved. A 
small injection of vasopressin (or re- 
lated small peptides) can protect the 


memory from puromycin blockage. The the ribosomes from bacteria passed 

end of this story is not yet in sight. through the cell. The polysomes, too, 

Possibly, protein synthesis is necessary were disrupted, 
for the final consolidation of long-term The product of this small group can 

memory, but this process continues dur- be measured by the three books and 200 

ing the week or so after training when or more papers published. A better meas- 

the memory trace seems to be spreading ure can be found in a listing of signifi- 

throughout the brain. Unfortunately, cant findings: (1) Isotopic competition 

this hypothesis cannot be tested at pres- method. (2) Flow patterns of metabolism 

ent because prolonged inhibition of pro- in E. coli. (3) Feedback inhibition. (4) 

tein synthesis is lethal. Demonstration that nascent protein is at- 

Most of these experiments were car- tached to ribosomes. (5) Virus purifica- 

ried out at the University of Pennsyl- tion on DEAE columns. (6) Patterns of 

vania by the Flexners. Active experi- ribosome synthesis. (7) Proof that newly 

mental work was a minor effort of the formed RNA in bacteria is one-third 

Biophysics Section in the period 1966- messenger, two-thirds ribosomal precur- 

1974. Roughly 20 papers have been sor. (8) Agar column for nucleic acid 

published, and the results up to 1969 reactions. (9) Messenger RNA isolation, 

are summarized in Quarterly Review of (10) Application of nucleic acid inter- 

Biophysics, 1969. action to measure relatedness of species 

and degree of divergence. (11) Homology 

n i • of DNA in E. coli and lysogenic virus. 

(12) Detailed studies of divergence in 

This brief history has attempted to viruses, bacteria, algae, vertebrates, 
trace the main themes followed by the rodents, and primates. (13) Application 
section during its 30 years. The past of nucleic acid reactions to cancer virus 
was emphasized because its individual investigations. (14) Demonstration that 
events are largely forgotten by now and DNA remains identical but RNAs differ 
the significant findings are merged into in different tissues and stages of devel- 
the general story of molecular biology, opment. (15) Extensive studies of nucleic 
The recent work is not quite ripe for acid reaction kinetics. (16) Discovery of 
such historical treatment. Neither was repeated DNA. (17) Properties of re- 
it possible to do justice to the many peated DNA. (18) Involvement of cate- 
fascinating offshoots that have been fol- cholamines and peptides in learning and 
lowed. The development of methods and memory. 

techniques, too, has been ignored. For We are pleased to have participted 

example, the pressure cell was first built in this exciting period in the development 

to release phage from infected bacteria of biology. We believe that we did make 

before lysis; this was achieved elsewhere significant contributions and that, since 

a short time later. Our attempts failed some of us will carry on in different 

because the pressure cell was too effective places, our history is not complete. We 

and disrupted the phage. The same effi- are especially pleased to note the contri- 

ciency kept us from observing polysomes butions being made by 22 Fellows who 

in spite of hundreds of observations of received a part of their training with us. 


P. H. Abelson, 1946-1953 B. J. McCarthy, 1958-1964 

D. B. Cowie, 1946- D. E. Kohne, 1965-1971 
R. B. Roberts, 1947- N. R. Rice, 1968- 

E. T. Bolton, 1949-1966 B. H. Hoyer, 1968- 
R. J. Britten, 1951-1971 



D. Axelrod 
G. N. Cohen 
H. H. Darby 

E. H. Davidson 
W. R. Duryee 
J. B. Flexner 
L. B. Flexner 

A. I. Aronson 
R. J. Avery 
J. A. Boezi 
T. I. Bonner 

D. J. Brenner 
T. J. Byers 

J. A. Chiscon 

E. H. Creaser 
M. Davis 

J. D. Duerksen 
N. J. Grady 

E. Aldous 
R. E. Bresnahan 
M. Chamberlin 
J. H. Eames 
E. F. French 
L. Gazze 
J. M. George 
J. Gray 
W. Griffin 
D. Johnson 
M. L. Kelley 
S. Lohman 

Resident Associates and Senior Fellows 

A. Gelderman 
M. A. Martin 
F. T. McClure 
K. McQuillen 
M. Miranda 
I: Z. Roberts 
S. Spiegelman 
P. Szafranski 


Y. Kato 
E. S. Kempner 
J. J. Leahy 
H. M. LenhofT 
S. Lovtrup 
J. E. Midgley 

A. V. Rake 
R. A. Shleser 
N. A. Straus 

D. G. Wallace 
M. J. Waring 

Technical Assistants 

M. Melean 

B. D. North 
P. Paul 

L. Prager 
P. Roddy 
M. K. Sands 
A. Shirven 
J. Smith 

E. Stern 

N. van de Velde 
J. Young 


L. T. Aldrich, M. Bass, L. Beach, C. Brooks, K. D. Burrhus, M. Dodson, S. E. Forbush, 
S. R. Hart, A. Hojmann, D. E. James, A. T. Linde, I. S. Sacks, N. Shimizu, 

A. Snoke, and W. White 


C. Aguirre, J. Bannister, M. Casaverde, A. Cuyubamba, G. L. Davis, S. del Pozo, 

D. W. Evertson, R. C. Fletcher, A. Flores, L. Gelinas, A. A. Giesecke, T. E. Krogh, I. Kushiro, 

Y . Motoya, L. Ocola, H. Okada, G. Olajsson, J. E. Ramirez, A. Rodriquez B., 

J. G. Schilling, H . Sigtryggsson, D. Simoni, R. Stefansson, S. Suyehiro, L. Tamayo, 

and Y. Yamagishi 

One of the goals in understanding the and interior. Over the past 30 years 

earth and its history is to learn about members of this Department have played 

the physical and chemical properties of a major role in initiating and advancing 

rocks that characterize the various prov- research on the isotopic composition and 

inces and domains of the earth's surface age of rocks and on the seismic structure 


of the crust and uppermost mantle. With they yield to these stresses. Reports by 

every advance in knowledge has come Bass, by Forbush, and by Aldrich and 

deeper appreciation of the earth's com- colleagues pertain to areas of geochemis- 

plexity and the scale of its inhomogeneity. try and geophysics other than those just 

With passing time, the studies of the outlined and emphasize the personal and 

Department centered increasingly on in- diverse research interests of the members 

vestigating the differences between var- of the Department. 

ious regions of the globe with the aim of Research by Hofmann and by Shimizu 

answering the question "Is there struc- and Kushiro provides fundamental in- 

ture and pattern to the complexities we formation on the way in which trace 

observe?" elements distribute themselves in the 

The theories of sea-floor spreading and mantle. Hofmann's report gives kinetic 
plate tectonics have provided a concep- data on rates of diffusion of Sr and Ca 
tual framework in which the history and that bear directly upon the question of 
processes of crust and mantle can be the degree of trace element homogeniza- 
better understood. As a result, ideas and tion one would expect between mineral 
concepts have evolved rapidly over the phases in a partially melted mantle, 
past several years. Of special importance Shimizu and Kushiro's report is another 
has been the restructuring of earth sci- in a continuing series demonstrating how 
ences as once widely separated disci- certain trace elements and rare earth 
plines in geology, geochemistry, and geo- elements are partitioned between crystal 
physics have merged to provide more and liquid phases during partial melting 
unified approaches to resolving long- of mantle material. Knowledge of par- 
standing problems. titioning coefficients provides a means of 

The blurring of distinctions between inferring mantle compositions from the 
geochemistry and geophysics has been compositions of the derivative magmas, 
evident over the past year in this De- Dodson's report on the kinematics of 
partment as our attention has focused metamorphic reactions provides further 
on the problem of the structure and com- insight into the rates at which mineral- 
position of the earth's mantle. Staff ogical changes occur in rocks with chang- 
Members and Fellows with widely diver- ing thermal regime, 
gent backgrounds, but drawing on each Reports by Hofmann and Hart, by 
other's expertise, are joined in several Hart, and by White, Hart, and Schilling 
common efforts to understand mantle deal with the problem of the origin of 
inhomogeneity and how it bears on plate mid-ocean ridge and oceanic island 
dynamics, the formation of magmas, and basalts. The two groups of volcanic rocks 
the evolution of crust and mantle. exhibit distinct differences in trace ele- 

Most of the geochemical studies re- ment and isotopic composition. Various 
ported here are concerned with questions and conflicting theories have been pro- 
of chemical and isotopic variation in posed to explain the differences, and the 
oceanic basalts and island arc andesites controversy now has resolved itself into 
as indicators of gross mantle inhomo- two contrary views: One holds that the 
geneity. Studies in geophysics span a compositional differences are due to dis- 
range of topics from upper mantle struc- equilibrium melting of a homogeneous 
ture in subduction zones to theoretical mantle in which mineral phases are not 
investigations of earthquake source in isotopic equilibrium with one another; 
models. A common goal of the studies the other holds that the magmas of 
in geophysics is to determine the physi- oceanic islands are derived from mantle 
cal properties of rocks within the crust of composition fundamentally different 
and mantle, the stresses to which they from that which supplies the mid-ocean 
are subjected, and the manner in which ridges. The reports presented here pro- 


vide strong support for the latter view, the downgoing slab in regions where the 
The authors support the idea that cer- seismicity is ambiguous or nonexistent, 
tain oceanic islands overlie mantle Snoke has successfully resolved a 
plumes or diapirs and that the source major controversy over incompatible 
material of these magmas was therefore differences between various earthquake 
derived from deep within the mantle. source models. Source models are corn- 
Brooks addresses the question of the monly used to infer fault dimensions and 
variation of mantle composition with stress release of an earthquake from the 
time. His studies on magma immiscibil- measured far-field radiation spectrum, 
ity, showing it to be a rare phenomenon Critical to any such study is whether the 
today but common in ancient rocks, sug- model is consistent with observation. A 
gest that both the composition of the model proposed by Archambeau ap- 
earth's mantle and its thermal regime peared to be in better argreement with 
may have been significantly different in some observations than are other source 
Archean time. models. Snoke shows, however, that the 
James, Brooks, and Cuyubamba carry effects predicted by Archambeau's model 
the study of mantle inhomogeneity to are due to nonphysical behavior and are 
the continents. They show that for the therefore spurious. Archambeau's model 
central Andean volcanic arc, the magmas is shown to be equivalent to other source 
of Mesozoic time were derived from a models. Linde, Sacks, and Snoke seek to 
mantle source of fundamentally different answer the question of why the observed 
composition from that which supplied spectra from earthquake radiation differ 
the magmas of late Cenozoic time, from those predicted by models. They 
James, in a discussion of a model for dis- show that multiple rupture (all source 
equilibrium melting of the thick ancient models assume simple rupture) during 
continental lithosphere of western South an earthquake can produce some of the 
America, shows that the high 87 Sr/ 86 Sr spectral effects observed. This in turn 
ratios of the late Cenozoic andesites and can lead to serious errors in the calcula- 
dacites could be due to varying degrees tion of apparent stress release and fault 
of partial melting of isotopically un- dimensions. 

equilibrated phlogopite in rigid litho- The discovery by workers in the Soviet 

spheric mantle. Union and in the United States that 

Sacks' investigations of upper mantle earthquakes are preceded by a long 

structure beneath the Japan arc shed period of dilatancy in the region of the 

light on how thick continental litho- eventual rupture has given added im- 

spheric plates may be formed. He shows petus to borehole strainmeter studies, 

that rigidity of the asthenosphere is ab- Sacks, Snoke, and others report on the 

normally high in that wedge of mantle sensitivity of the borehole strainmeters 

overlying the descending plate and situ- to the detection of precursory dilatancy 

ated between trench and volcanic arc. and calculate the distances from the 

He hypothesizes that in this zone, high fault at which dilatancy could be ob- 

rigidity asthenosphere is being "frozen" served for different fault lengths and 

out of normal asthenosphere and added noise levels. 

to the base of the island arc lithosphere. Aldrich and colleagues report on the 
This process would create a lithospheric results of continuing studies of the elec- 
"root" similar to that observed beneath trical conductivity anomaly beneath the 
South American and other continental Andes of South America. They have ex- 
areas. Snoke and Sacks in a companion tended their investigations into Chile 
paper on structure in subduction zones and Argentina and have obtained more 
show how converted and reflected seismic detailed observations in Peru and Bo- 
phases can be used to map the top of livia. Preliminary results indicate that 



the electrical conductivity structure be- 
neath southern Peru and Bolivia is com- 
plex and distinctly different from the 
conductivity structure observed in cen- 
tral Chile. 

The reports end with that of Forbush 
and Beach who demonstrate that cosmic- 
ray diurnal anisotropy is due to reversals 
in the sun's poloidal magnetic field. With 
this demonstration they are able to con- 
clude that all variations in cosmic-ray 
intensity, except those of meteorological 
cause, are due to solar mechanisms. 

The studies described in the reports 
that follow reflect the highly personal 
and individual approach to research at 
the DTM and testify to the soundness 
of Merle Tuve's proposition that a small 
group of scientists, working together in 
shared interest, pursuing their individual 
goals, can make significant contributions 
to knowledge. 

Diffusion of Ca and Sr in a 
Basalt Melt 

Albrecht W . Hofmann 

Interest in diffusion measurements has 
recently increased because of the current 
controversy in the literature over the 
validity of using chemical and isotopic 
differences in oceanic basalts as indi- 
cators of compositional differences in the 
source mantle (e.g., Hart et al., 1973; 
O'Hara, 1973, 1975; O'Nions and Pank- 
hurst, 1974; Flower et al, 1975). Those 
who deny the validity of these indicators 
assume that the source is homogeneous 
on a large scale and explain all chemical 
and isotopic variation found in the vol-' 
canic rocks as the result of differences in 
the melting and crystallization process 
(for a more detailed discussion see the 
companion reports by Hofmann and 
Hart and by James). All parties to the 
argument agree that partial melting in 
an initially homogeneous mantle and 
subtraction of the melt (by intrusion) 
leads to an inhomogeneous residue. If 
the mantle convects, as it must if the 
present concepts of plate tectonics are 

approximately correct, it seems likely 
that the inhomogeneous residue will be 
transported to a location where melt- 
ing occurs again (for a contrary view, 
see Ringwood, 1974). It is therefore im- 
portant to determine whether and on 
what scale the mantle can be rehomo- 
genized by the combined process of con- 
vection and diffusion. The results re- 
ported below confirm the prediction 
made in last year's Report that diffusion 
distances in a partially molten mantle 
are limited to tens or hundreds of meters 
and that rehomogenization of larger 
scale inhomogeneities (tens of kilo- 
meters) is highly unlikely. 

Diffusion data are also important in 
determining the structure of silicate 
melts and in understanding crystalliza- 
tion processes. However, a considerably 
larger body of data will be needed before 
applications in these areas can be real- 
ized. The present report is a continuation 
and expansion of work started last year 
(see Year Book 73, pp. 935-941). 

Experimental Methods 

Only aspects of the methods will be 
described which differ from those re- 
ported last year (Hofmann, Year Book 
73, pp. 935-941). Continued difficulties 
in obtaining leak-free and bubble-free 
diffusion couples resulted in several 
changes in technique: (1) In most of the 
experiments the radioactive tracer was 
applied in the form of a thin source at 
the open end of a platinum tube filled 
with quenched basalt glass (olivine tho- 
leiite flow from Kilauea, 1921), rather 
than in the form of an infinitely extended 
source (referred to as diffusion couple). 
(2) Most of the experiments were made 
using the beta emitter 45 Ca and the re- 
sults were analyzed by autoradiography 
rather than particle counting. This per- 
mitted a rough evaluation of the con- 
vective contribution to the total trans- 
port. (3) A new well-type Nal (3 inch 
diameter, 1X2 inch well) afforded 
improved resolution and counting geom- 
etry for gamma-ray spectrometry of 


85 Sr-doped charges. This makes it pos- posed for several days to a nuclear- 

sible to analyze the activity profile of emulsion plate (Ilford emulsion type 

85 Sr in the basalt-filled tubes by count- K.5, 25 m/x thick), and the density of 

ing either the sectioned material itself the developed image was scanned using a 

or the residual activity in the tube after microdensitometer with chart recorder 

grinding. output. Two typical records are shown 

The results obtained by the thin- in Fig. 39. 

source technique, though different in Charges doped with 85 Sr were analyzed 

some details, confirm the order of mag- by consecutive sectioning in a manner 

nitude of the results obtained by the similar to that described in Year Book 

diffusion-couple technique. This is en- 73, pp. 935-941. Improvements were 

couraging in view of the fact that meas- made by using the above-mentioned new 

ured diffusion coefficients in solid sili- detector crystal and by employing the 

cates frequently differ by several orders residual activity directly in the data 

of magnitude, depending on the experi- analysis (Frischat and Oel, 1966). 

mental technique (see, for example, One advantage of the improved ana- 

Giletti, 1974). lytical methods is that deviations of the 

Thin-source diffusion runs were pre- measured activity profile from an ideal 

pared as follows: Two-mm Pt tubing was diffusion profile are easily detected. Such 

filled with olivine tholeiite basalt melt deviations must be attributed to trans- 

by capillary action or by applying sue- port mechanisms other than volume 

tion at about 1400°O. The quenched diffusion, because in tracer diffusion any 

tubes were then cut into pieces 0.5 to concentration dependence of the diffusion 

2.5 cm long and reannealed at the ap- coefficient is assumed to be negligible, 
proximate intended run temperature. In 

the temperature range employed (1250° Results 

to 1450°C), the melt is held in the hori- Typical activity profiles are shown in 

zontally suspended, open-ended tubes by Figs 39 and 40 for thin-source diffusion 

surface tension. A final anneal for about of 45 Ca and 85 Sr> respectively. Figure 39 

1 minute at 1000°C caused the glass to also shows the deviations from "ideality" 

devitnfy and shrink so that the end encountered in diffusion runs of short 

surfaces were concave and had a dull duration. The comparison shows that a 

appearance. Several drops of 45 Ca or transport process other than volume dif- 

85 Sr chloride solution were then dried fusion is dominant during the early 

on one or both free ends of basalt, and sta ge S f a run. Some convection near 

the prepared charge was suspended at the open ends of the tube is likely during 

above-hquidus temperatures in a Pt- the initial heating of the charge. In 

wound furnace in air for durations of up addition, the distribution of tracer is 

to several days. This produced diffusion une ven (Fig. 41a). In order to minimize 

gradients of up to 1.5 cm long. t he effects of poorly controlled initial 

^ To determine the activity profile of and boundary conditions, a study of the 

Ca (/? , 0.255 MeV, t 1/2 = 165 days) time dependence of the transport dis- 

m the basalt after the high-temperature tance was made at 1310 ± 10°C. The 

diffusion experiment, the glass-filled Pt reS ults, in terms of apparent diffusion 

tube was mounted with epoxy cement in coefficients, are shown in Fig. 42. This 

a groove on the surface of an aluminum an d the much closer correspondence of 

block. This assembly was hand ground long-run profiles to "ideal" diffusion 

and polished to remove the upper half profiles (Fig. 39) indicate clearly that 

of the tube and to expose a flat, polished volume diffusion becomes the dominant 

section parallel to the long axis of the transport process in runs lasting several 

tube. The polished block was then ex- hours or days. 



Run 74-32 
T = 1309 °C 

t= 1.21 hr 

Run AH75-9 
t = 47.75 h 


Fig. 39. 45 Ca activity in basalt as a function of distance in two thin-source experiments. The 
activity is determined by microdensitometry from a nuclear emulsion. The solid lines are 
calculated diffusion profiles for diffusion from an infinitely thin source into a semi-infinite 
medium. Run durations (and temperatures) are 12.1 hours at 1309 °C (a) and 47.75 hours at 
1390°C (6). Similar deviations between measured profile and the calculated line are evident 
in all runs of short duration. 



Length (cm) 

Fig. 40. Residual ^Sr activity versus length determined by successive removal of basalt 
glass and counting of the remaining basalt rod. The line is a calculated curve for ideal thin- 
source diffusion into a semi-infinite medium. 

Fig. 41. Images of beta particle tracks on nuclear emulsions: a, taken from a 89-hr diffusion 
run at 1309°C ; b, taken from a 3.8-hr diffusion run at 1305°C. 



t (secxICT 4 ) 

Fig. 42. Apparent diffusion coefficients at 1310 ± 10 °C as a function of run duration, deter- 
mined from two points in each activity profile. These are measured at A/A = 0.1 (crosses) 
and A/A = 0.5 (circles), where A is the activity at the origin. The percent difference between 
the two apparent D values is an indication of the misfit between the measured profile and an 
ideal diffusion profile from a thin source. This difference decreases from about 100% for short 
times to less than 10% for long run times. 

The temperature dependence of the 
diffusion coefficient for 45 Ca is shown in 
Fig. 43, which also gives the results of 
Medford (1973) on Ca diffusion in 
mugearite for comparison. The differ- 
ences between the two sets of data may 
be because of the difference in matrix 
composition of the two melts employed; 
or it may be because Medford measured 
chemical interdiffusion of calcium with 
other components of the basalt, whereas 
the present experiments were designed 
to measure Ca-tracer diffusion. 

Figure 44 shows the results of the 
experiments for 85 Sr diffusion. In this 
case there is a discrepancy between the 
results obtained by the couple method 
(as described in last year's Report) and 
the new thin-source data. It had been 
estimated (Hofmann, Year Book, 73, p. 
938) that the preliminary data on dif- 
fusion couples were accurate within a 
factor of 2, so the new results may not 
differ significantly from the previous 
ones. Additional experimentation is 
clearly needed to obtain a better estimate 

of the temperature dependence and to 
demonstrate that the two techniques 
yield concordant results. 

The order of magnitude of the diffu- 
sion coefficients of both Sr and Ca is 
such that equilibration of the melt phase 
on a centimeter scale in a partially 
molten mantle requires only a few years, 
but equilibration on a kilometer scale 
requires more than one lifetime of the 
earth. Diffusion in solid silicates tends 
to be slower by several orders of mag- 
nitude (see Hofmann and Hart, this 
Report), so that equilibration in an all- 
solid mantle is restricted to an even 
smaller scale. 


The high-temperature experiments 
were done at the Geophysical Laboratory 
with help and advice from the staff and 
Fellows. The basalt sample was supplied 
by H. S. Yoder, Jr. Financial support 
was given by the National Science 
Foundation, Grant no. GA-40250. 




I0~ 6 


1460 1420 





1 1 1 1 



1 1 1 








I i 

i i 


1 1 






I0 4 /T(°K _I ) 

Fig. 43. Arrhenius plot (log D versus \/T) for ^Ca-tracer diffusion in olivine tholeiite melt. 
The line a is the (visual) best fit through the measured values of the diffusion coefficient. Its 
slope corresponds to an activation energy of 44 kcal/mole. Line b is taken from Medford 
(1973) and represents Ca diffusion in a mugearite melt. 

T (°C) 

I0 _b — 





1420 1380 




1 1 

1 1 1 


1 1 

1 1 





\ O 











Cs \ 


1 1 


1 \| 








I0 4 /T(°K _I ) 

Fig. 44. Arrhenius plot for ^Sr-tracer diffusion in olivine tholeiite melt. Crosses and circles 
represent D values from diffusion couples and thin-source experiments, respectively. The line 
labeled Ca represents ^Ca-tracer diffusion and is taken from Fig. 43. The line labeled Cs repre- 
sents Cs-tracer diffusion in a synthetic basalt-like melt determined by Winchell and Norman 


Flower, M. F. J., H. U. Schmincke, and 
R. N. Thompson, Phlogopite stability 
and the 87 Sr/ 86 Sr step in basalts along 
the Reykjanes Ridge, Nature, 25 '4, 
404-406, 1975. 

Frischat, G. H., and H. J. Oel, Eine 
Restaktivierungsmethode zur Bestim- 

mung von Selbst diffusions — Koeffi- 
zienten in Festkorpern, Z. Angew. 
Phys., W, 195-201, 1966. 
Giletti, B. J., Diffusion related to geo- 
chronology, in Geochemical Transport 
and Kinetics, A. W. Hofmann, B. J. 
Giletti, H. S. Yoder, Jr., and R. A. 
Yund, eds., Carnegie Inst. Wash. Publ. 
634, PP- 61-76, 1974. 



Hart, S. R., J. G. Schilling, and J. L. 
Powell, Basalts from Iceland and 
along the Reykjanes Ridge: Sr isotope 
geochemistry, Nature, 246, 104-107, 

Medford, G. A., Calcium diffusion in a 
mugearite melt, Can. J. Earth Sci., 
ID, 394-402, 1973. 

O'Hara, M. J., Non-primary magmas 
and dubious mantle plume beneath 
Iceland, Nature, 21+3, 507-508, 1973. 

O'Hara, M. J., Is there an Icelandic 
mantle plume?, Nature, 253, 708-710, 

O'Nions, R, K., and R. J. Pankhurst, 
Petrogenetic significance of isotope 
and trace element variations in vol- 
canic rocks from the mid- Atlantic, 
J. Geol, 15, 603-634, 1974. 

Ringwood, A. E., The petrological evo- 
lution of island arc systems, J. Geol. 
Soc. London, 130, 183-204, 1974. 

Winchell, P., and J. H. Norman, A study 
of the diffusion of radioactive nuclides 
in molten silicates at high tempera- 
tures, in High Temperature Technol- 
ogy, third International Symposium 
1967, 1969. 

The Partitioning of Rare Earth 

Elements Between Garnet and Liquid 

at High Pressures : Preliminary 


N. Shimizu and I. Kushiro 

The interpretation of the trace ele- 
ment abundances of volcanic rocks in 
terms of fractional crystallization and/ 
or partial melting models rests upon a 
knowledge of the partitioning of the 
trace elements between liquid and solid 
phases in these processes. The partition- 
ing of rare earth elements (REE) be- 
tween garnet and liquid is particularly 
important for explaining the enrichment 
of light REE relative to heavy REE in 
alkali basalts and kimberlites as these 
light REE enriched patterns have been 
ascribed to equilibration with garnet 
(e.g., Gast, 1968; Schilling and Win- 
chester, 1969; Kay and Gast, 1973). 

Experimental evidence has been needed 
to show the applicability of the partition 
coefficients of REE measured for natural 
garnet-liquid pairs (Schnetzler and Phil- 
potts, 1970; Philpotts et al, 1972) to 
partial melting of garnet lherzolite or 
crystallization of garnet of the liquidus 
of basaltic magmas at high pressures. 

The composition chosen for the study 
is diopside 30% (by weight), pyrope 
70%. The mixture of this composition 
studied by O'Hara and Yoder (1967) is 
used here as the starting material. A 
composite solution containing eight REE 
was added to the powdered material in 
a Pt crucible and dried on a hot plate. 
It was then fused at 1500°C in air to 
make a homogeneous glass starting ma- 
terial for high-pressure runs. Concen- 
trations of REE in the starting material 
are given in Table 6 and are about 15 
to 40 times chondritic average. They are 
within the range of concentrations ob- 
served for natural basalts. 15 to 20 mg 
of the powdered material were sealed 
into platinum capsules with 3.5 to 10.2 
wt% of water. Ten runs were carried 
out at 30 kb at temperatures between 
1175° and 1450°C for 2 hours in a solid- 
media high-pressure apparatus. The best 
run product was obtained at 1275°C 
with about 9 wt% water. It consisted of 
euhedral garnet crystals 10 to 30 /un 
in diameter (about 25% of the charge), 
clear glass, quench crystal (dendritic or 
feathery), and a trace amount of forste- 
rite. In order to evaluate equilibrium 
relationships between garnet and liquid 
(glass and quench crystal), the major 
element composition of garnet was de- 
termined by an electron microprobe. The 
garnet composition (Si0 2 , 44.5; A1 2 3 , 
23.1; MgO, 27.1; CaO, 3.94; total, 
98.64) was found to be quite homogene- 
ous from one grain to another, indicating 
that equilibrium was closely approached 
at least for the major elements. 

The run product was then pulverized 
with two-bottle distilled acetone in an 
agate mortar and processed with the 
differential dissolution technique (DDT) 


TABLE 6. REE Concentrations and Partition Coefficients (D) 


LIQ (ppm)* 


SM (ppm)* 









































* GAR, garnet; LIQ, liquid; SM, starting material. 
f Plotted as curve A in Fig. 46. 

developed (Shimizu and Hart, 1973) for crystal together with glass. The presence 
separation of phases in a similar experi- of a trace amount of forsterite is, how- 
mental study of the partitioning of LIL ever, believed to be of very minor im- 
trace elements between clinopyroxene portance. During the course of the DDT 
and liquid at high pressures (Shimizu, treatments, it was observed that a part 
1974). Rare earth elements (REE) were of the quench crystals was hard to dis- 
analyzed in the separated phases by solve in dilute HF (5% for example) and 
isotope dilution using spike solutions that a part of the garnet was dissolved 
provided by Dr. J. A. Philpotts. Total after several steps of treatment with 
blanks were 0.1 ng for Ce and 0.01 to 15% HF. When some part of the liquid 
0.06 ng for the other REE and were phase remains after the HF treatments, 
negligible except for Ce in garnet for it causes an anomalously high concentra- 
which a correction of about 12% was tion of light REE in the solid phase in 
required. We consider that the concen- a manner similar to a matrix impurity 
trations reported in Table 6 are reliable in a mineral separate of phenocryst- 
to 5%. matrix pairs and thus increases the ap- 
In this preliminary study, the vari- parent partition coefficients. The effect 
ation of partition coefficients with P, T, can be significant if the true equilibrium 
and composition was not investigated, partition coefficients are small (<1). 
Several runs were carried out at identi- For heavy REE, on the contrary, the 
cal P and T (30 kb, 1275°C) and almost presence of quench crystal impurity 
identical water content (about 9 wt%) would decrease the concentrations and 
to test the DDT with different concen- thus lower the apparent partition co- 
trations of HF. The concentrations of efficients. The magnitude of the impurity 
REE in the liquid phase (glass and effect may be estimated by simple cal- 
quench crystal) were obtained from the culations. If the true partition coefficient 
first-step dissolution as in the previous for Ce is 0.01, presence of 1% liquid 
study, and those in garnet were obtained phase would increase the concentration 
for the sample remaining after five con- by a factor of 2 and thus increase the 
secutive HF treatments. Although the partition coefficient by the same magni- 
separation technique (DDT) was found tude. In contrast, the same amount of 
to be excellent in the previous study on liquid phase would decrease the Yb con- 
clinopyroxene-liquid (Shimizu, 1974), in tent of garnet only by 0.7% and thus 
which only clinopyroxene and clear glass have no significant effect, if the true 
were observed, a serious difficulty was partition coefficient for Yb is around 4. 
encountered in the present experiments Comparison of the several runs (two 
in part due to the presence of quench of them are shown in Fig. 45 as curves 



A and B) suggests that this is the case concentration of the acid in the present 

in the present experiments. Curve A was experiments). Curve B is an example of 

obtained for the run product which con- a less successful separation in part due 

tained the smallest amount of quench to an abundance of quench crystal and 

crystal and which was treated with 10% to use of dilute HF (5%). It should be 

HF (which seemed to be an optimum noted that these curves appear to be 















0.1 - 

0.01 - 


i i i r i i 

I I I I I I I 


,~~" 218 











i ^ - 




/ / - 


/ Ai 

/ /■> E 


/ Y 

/.-yT I 

_ / 



— *- ** * 

r x 


' sfc 


,<' // 



'- / / / 


■ // 


/ //' 

/ V 


/ // 

/ // 

/ // 218 


"A® // 

// 21 f 

Philpotts et al. 

. // 

; / 


I I I I I I 




Sm Eu Gd 




Fig. 45. Partition coefficients of REE between garnet and liquid. A and B, this study (see 
text) ; 218, Schnetzler and Philpotts (1970) ; 21f, Philpotts et al. (1972) ; E, estimate (Shimizu, 


rotating around Dy for which the par- example of the estimated partition co- 
tition coefficient is close to 1. Although efficients obtained in this manner. The 
the values for A seem to be best in the experimental values are in excellent 
present experiments, we still are not sure agreement with the estimated ones, 
that the phase separation was complete. It is clear from Fig. 45 that REE of 
Therefore, we suggest that the partition larger ionic radii are rejected more 
coefficients for the light REE, particu- strongly from garnet (relative to liquid), 
larly that for Ce, should be regarded as This relationship is better illustrated in 
upper limits. The partition coefficient for Fig. 46, where the partition coefficients 
Ce may have to be reduced by a factor (now defined as (REE/Mg) G Ait/(REE/ 
of 2 (to 0.01) if the final purity of the Mg) LIQ , plotted as the natural loga- 
garnet phase is 99%. Nevertheless, we rithm) are plotted against the ionic 
suggest that the values for the heavy radii (values from Whittaker and Mun- 
REE, which are of particular importance tus, 1970: 8-fold coordination, assuming 
for the REE geochemistry of volcanic that REE substitute for Mg by analogy 
rocks, are reliable, since the effect of with Y-garnet). The present results de- 
impurity seems to be small, as men- fine a straight line for REE ranging 
tioned above. from Yb to Nd, while the point for Ce 

In Fig. 45 the partition coefficients is significantly lower. The major ele- 

obtained for natural garnet-liquid pairs ment composition of the liquid phase was 

are plotted for comparison. The values calculated by material balance, based 

for 218 (Schnetzler and Philpotts, 1970) on garnet composition (probe data) and 

are for the almandine-rich garnet pheno- starting material and on proportions of 

cryst-dacitic matrix pair. They are sig- garnet and liquid estimated from REE 

nificantly higher than the experimental material balance (26% garnet and 74% 

values for all REE. Although the causes liquid; trace amount of forsterite was 

of these differences are not well under- neglected). 

stood, it should be noted that the chemi- It is noted that the partition coeffi- 

cal compositions as well as P and T are cients for Yb and Er are larger than that 

grossly different between the almandine- of Mg despite their considerably larger 

dacite pair and the present experiments, ionic radii. This situation appears to be 

Better agreement is obtained for a garnet similar to the plagioclase-liquid system, 

megacryst-host basalt from Kakanui, in which the partition coefficient of Sr is 

New Zealand. higher than that of Ca, Sr having a 

In recent papers (e.g., Allegre et al., larger ionic radius than Ca (Higuchi and 
1973; Kay and Gast, 1973; Frey et al., Nagasawa, 1969). It has been suggested 
1974; Shimizu, 1975), the partition co- that Sr would be of a more suitable size 
efficients for GAR-LIQ are estimated in the plagioclase structure than Ca 
from CPX-LIQ and GAR-CPX rela- (Higuchi and Nagasawa, 1969). By 
tionships. It has been shown (Philpotts analogy, it may be suggested that the 
et al., 1972; Shimizu, 1975) that the heavy REE (Yb and Er) are more suit- 
partitioning of REE between garnet and able elements in the garnet structure 
clinopyroxene is rather constant over a than Mg. 

range of chemical composition and esti- Considering that the partition coeffi- 

mated P and T. This seems to justify cient for Ce is an upper limit, the devi- 

combining the natural CPX-GAR rela- ation of Ce from the straight line (Fig. 

tionships and CPX-LIQ partitioning 46) seems significant. The deviation is 

(particularly those of Grutzeck et al., related to the ionic radius in such a way 

1973, or Schnetzler and Philpotts, 1970) that for ions larger than Ca (the radius 

to estimate GAR-LIQ partition coeffi- of Ca = 1.20A), being equal to that of 

cients. Curve E in Fig. 45 shows an Nd), the energetics of substitution may 




I I 

1 1 1 



1 1 

1 l 

1 1 1 


D= ( 

REe/m 9 ) 



REE /H,a 



• \ 



— • 






i i i 









1 I 






Nd 1 \ 
Ca Ce x 

i i i 


1.00 1.10 1.20 

ionic radius (A) 

Fig. 46. InD vs ionic radii (8-fold coordination, from Whittaker and Muntus (1970)) 

be different from that for the smaller 
REE. The estimated partition coefficient 
for La (Frey et al, 1974) of 0.004 is on 
the extension of the Nd-Ce line. The 
higher oxidation state of Ce 4+ as a rea- 
son for the deviation mentioned above 
may be ruled out because Ce 4 + has a 
smaller ionic radius than Ce 3 + . 

The arguments given above would be 
valid only for pyrope-rich garnet. For 
grossularite-andradite garnet, the sites 
available for REE are larger because Ca 
is the host ion. It would be possible then 
to expect that quite a different partition 
pattern would be observed for the gros- 
sularite-liquid system than that of the 
present experiments. 

Agreement between the present results 
and the garnet-liquid partition coeffi- 
cients estimated for natural systems sug- 
gests that these values can be used to 
calculate the REE patterns of liquids 
produced by partial melting of garnet 
lherzolite and to assess the effect of 
"eclogite fractionation" (O'Hara and 
Yoder, 1967). An example of this appli- 
cation is found in Shimizu and Arculus 

(1975), who examined partial melting 
and fractional crystallization models 
(including eclogite fractionation) based 
on the partition coefficients reported here 
for the REE variations observed for a 
suite of basanitoids-alkali olivine basalts 
from Grenada, Lesser Antilles. They 
concluded that the observed variations 
(a factor of 6 for Ce and constant Yb) 
may be best explained by various degrees 
of partial melting of garnet lherzolite. 
O'Hara (1968) stated that the higher 
K 2 content of alkali basalt compared 
to tholeiite may be explained by a small 
degree of partial melting (5% for alkali 
basalt and 20% for tholeiite) followed 
by extensive eclogite fractionation (50%, 
compared with no eclogite fractionation 
for tholeiite) and subsequent olivine 
fractionation (40%, same for tholeiite), 
starting with the mantle with K 2 as 
low as 0.025%. If these figures are to be 
taken as representative, relatively small 
variations in heavy REE (8-15 times 
chondrite) between tholeiite and alkali 
basalts suggests that a high partition 
coefficient of heavy REE (Yb) between 



garnet and liquid (such as 40, Schnetzler 
and Philpotts, 1970) is difficult to recon- 
cile with the observation. For example, 
50% eclogite fractionation could result 
in depletion of Yb in the residual liquid 
of more than 4 orders of magnitude if 
j)Yb __ 4Q anc j ec i mte consists of 50% 

Gar ° 

garnet and 50% clinopyroxene. 

O'Hara (1970) suggested that the 
chemical composition of partial melts 
and solid residua can be satisfactorily 
explained by garnet lherzolite-garnet 
harzburgite inclusions in kimberlite 
(GLIK/GHR model). This implies that 
a significant amount of garnet is present 
in the solid residua (in fact, some garnet 
harzburgite inclusions contain more than 
5% garnet) . If this were the typical cir- 
cumstance for partial melting of the 
mantle, and if the REE patterns of 
basaltic magmas represent partial melts, 
then calculations suggest that a very 
high partition coefficient of heavy REE 
(Yb) between garnet and liquid (such as 
40) is difficult to reconcile with the 
petrologic models. For example, a liquid 
in equilibrium with a solid containing 
garnet (5% of the original solid, for 
example) would be depleted in Yb by a 
factor of about 2 relative to the initial 
solid. If the mantle has REE abundances 
of 2-4 times chondrite, the above liquid 
must have Yb content of 1-2 times 
chondrite, being much too low for any 
erupted magmas. 


Allegre, C. J., R. Montigny, and Y. Bot- 
tinga, Cortege ophiolitique et cortege 
oceanique, geochimie comparee et 
mode de genese., Bull. Soc. Geol. Franc., 
15, 461-477, 1973. 

Frey, F. A., W. B. Bryan, and G. Thomp- 
son, Atlantic ocean floor: geochemistry 
and petrology of basalts from Legs 2 
and 3 of the Deep Sea Drilling Proj- 
ect, J. Geophys. Res., 79, 5507-5527, 

Gast, P. W., Trace element fractionation 
and the origin of tholeiitic and alka- 

line magma types, Geochim. Cosmo- 
chim. Acta., 32, 1057-1086, 1968. 

Grutzeck, M. W., S. J. Kridelbaugh, and 
D. F. Weil, REE partitioning between 
diopside and silicate liquid, Eos, 54, 
1222, 1973. 

Higuchi, H., and H. Nagasawa, Parti- 
tion of trace elements between rock- 
forming minerals and the host volcanic 
rocks, Earth Planet. Set. Lett., 7, 281- 
287, 1969. 

Kay, R. W., and P. W. Gast, The rare 
earth content and origin of alkali-rich 
basalts, J. Geol, 81, 653-682, 1973. 

O'Hara, M. J., The bearing of phase 
equilibria studies in synthetic and nat- 
ural systems on the origin and evolu- 
tion of basic and ultrabasic rocks, 
Earth Sci. Rev., 4, 69-133, 1968. 

O'Hara, M. J., Upper mantle composi- 
tion inferred from laboratory experi- 
ments and observations of volcanic 
products, Phys. Earth Planet. Int., 3, 
236-245, 1970. 

O'Hara, M. J., and H. S. Yoder, Jr., 
Formation and fractionation of basic 
magmas at high pressures, Scott. J. 
Geol, 3, 67-117, 1967. 

Philpotts, J. A., C. C. Schnetzler, and 
H. H. Thomas, Petrogenetic implica- 
tions of some new geochemical data 
on eclogitic and ultrabasic inclusions, 
Geochim. Cosmochim. Acta, 36, 1131— 
1166, 1972. 

Schilling, J. G., and J. W. Winchester, 
Rare earth contribution to the origin 
of Hawaiian lavas, Contrib. Mineral. 
Petrol, 23, 27-37, 1969. 

Schnetzler, C. C., and J. A. Philpotts, 
Partition coefficients of rare earth ele- 
ments between igneous matrix material 
and rock-forming mineral phenocryst, 
II, Geochim. Cosmochim. Acta, 34, 
331-340, 1970. 

Shimizu, N., An experimental study of 
the partitioning of K, Rb, Cs, Sr and 
Ba between clinopyroxene and liquid 
at high pressures, Geochim. Cosmo- 
chim. Acta, 38, 1789-1798, 1974. 

Shimizu, N., Rare earth elements in gar- 
nets and clinopyroxenes from garnet 



lherzolite inclusions in kimberlites, 
Earth Planet. Sci. Lett., 25, 26-32, 

Shimizu, N., and R. J. Arculus, Rare 
earth element concentrations in a 
suite of basanitoids and alkali olivine 
basalts from Grenada, Lesser Antilles, 
Contrib. Mineral. Petrol., in press, 

Shimizu, N., and S. R. Hart, Differential 
dissolution technique (DDT) : chemi- 
cal separation of crystals from glass, 
in Carnegie Inst. Wash. Year Book 72, 
pp. 268-270, 1973. 

Whittaker, E. J. W., and R, Muntus, 
Ionic radii for use in geochemistry, 
Geochim. Cosmochim. Acta, 34, 945- 
956, 1970. 

An Assessment of Local and Regional 

Isotopic Equilibrium in a Partially 

Molten Mantle 

A. W. Hojmann and S. R. Hart 


The isotopic composition of Sr and Pb 
has been used to set constraints on the 
origin and the composition of the source 
material of volcanic rocks. Gast (1960) 
pioneered this approach and showed 
by comparing meteoritic and terrestrial 
87 Sr/ 86 Sr ratios and trace element abun- 
dances that the earth (or at least the 
upper mantle and crust) is depleted in 
alkalies relative to chondritic meteorites. 
Starting with the work of Hurley et al. 
(1962) and Faure and Hurley (1963), 
the initial 87 Sr/ 86 Sr ratio was used to 
determine, for example, whether granites 
were derived from the crust or the man- 
tle. Gast, Tilton, and Hedge (1964) for 
the first time showed that there are con- 
sistent differences between individual 
oceanic islands in the isotopic composi- 
tion of lead and strontium, and they 
concluded that there are regional varia- 
tions in the composition of the upper 
mantle. Hedge and Peterman (1970) and 
Hart (1971) compared 87 Sr/ 86 Sr ratios 
from ocean floor basalts (i.e., rocks de- 
rived from mid-ocean ridges) with ratios 

from oceanic islands and showed that 
the isotopic ratios from the ocean floor 
are systematically lower. More recent 
work has shown that although signifi- 
cant regional differences do exist, the 
ocean floor basalts are the most nearly 
constant in their Sr-isotopic composition 
of all the volcanic rocks investigated. 
The 87 Sr/ 8(; Sr ratios of these rocks 
usually vary within the narrow range 
0.7025-0.7035, with the large majority 
having values less than 0.7030. In con- 
trast, the 87 Sr/ 86 Sr ratios of oceanic is- 
land basalts are higher and much more 
variable. The pattern that emerges from 
these results is one of distinct groupings 
of basalts from the different environ- 
ments in terms of isotopic compositions. 
This is illustrated in Fig. 47, where 
87 Sr/ 86 Sr ratios of basalts are shown for 
rocks derived from the ocean floor and 
from ocean islands. These results imply 
that the mantle is compositionally in- 
homogeneous both laterally and verti- 
cally because volcanoes situated near a 
mid-ocean ridge usually erupt lavas that 
differ in isotopic composition from the 
mid-ocean ridge material itself. If these 
conclusions are valid, they constrain the 
choice of convection patterns in the 
mantle. For example, models that call 
for derivation of all volcanic rocks from 
the same convecting upper mantle must 
be rejected. On the other hand, models 
that call for derivation of mid-ocean 
ridge material from a partially molten 
upper mantle and derivation of oceanic 
islands from plumes or blobs rising from 
deeper within the mantle are consistent 
with the isotope data (see also Schilling, 

Until recently the interpretation out- 
lined above has in general been accepted 
as valid by isotope geochemists and 
ignored by petrologists and geophysi- 
cists. Now, presumably for the purpose 
of circumventing the consequences of 
these restrictions, the basic assumption 
of all conclusions based on isotopic data 
has been challenged. This is the assump- 
tion that the isotopic composition of the 



Jan Mayen (1,2) 

Kolbeinsey ( I ) 

Iceland (postglaciol )( 1 ,3,4,5,6) 

Azores (1,7) 

Canary Islands (I ) 

Cape Verde (8) 

Ascension (I) 

St. Helena (9) 

Tristan da Cunha ( I ) 

Bouvet ( I ) 

Crozet (101 

Kerguelen (10) 

St. Paul (10) 

Amsterdam (10) 

Reunion (10) 

Eniwetok ( II ) 

Samoa (12 ) 

Hawaii (13) 

Easter Island (II ) 




1 2 6 |8|4|5~~[T 

41010 5 2 3 

cn a 


M6lfl4|4 | m 




□ □ a 

a a 


a m a 

irrn m nn 





Sr 87 /Sr 86 




Fig. 47. Comparison of 87 Sr/ 86 Sr ratios from mid-ocean ridge basalts (MORB) and from 
oceanic islands. The MORB data are plotted as a frequency diagram (number of samples (n) 
versus 87 Sr/ 86 Sr) below the horizontal line. The island data are arranged geographically in the 
sequence North Atlantic, South Atlantic, Indian, and Pacific Oceans. The number of samples 
for each value of 87 Sr/ 86 Sr (rounded to four significant figures) is indicated on the figure; a 
blank square represents one sample. The following data have been omitted: (1) all island-arc 
data; (2) all analyses published prior to 1970 (because of less uniform analytical precision of 
these older analyses) ; (3) analyses of samples from the Reyk janes Ridge between and 200 
km from Iceland (because this region is similar to Iceland itself with respect to isotopic and 
trace element composition), see reference 3 below; (4) one anomalously low value given by 
reference 17 (because reanalyses of the sample failed to confirm this value) ; (5) all data for 
mid-Atlantic ridge samples by White et al., this Report) (because the results are shown else- 
where in this Report, and because inclusion of these data would aggravate the problem 
inherent in this type of frequency plot, namely, weight the data excessively with a large num- 
ber of analyses from one area, the Azores platform). The key on the left-hand side identifies 
the islands and the references for islands and ocean-floor samples. The references are indicated 
by a number in parentheses and by the following list: (1) O'Nions and Pankhurst, 1974; 
(2) Lussiaa-Berdou-Polve and Vidal, 1973; (3) Hart et al, 1973; (4) Sun and Jahn, 1975; 
(5) O'Nions et al, 1973; (6) O'Nions and Gronvold, 1973; (7) White et al, 1975; (8) Klerkx 
et al, 1974; (9) Grant et al, 1974; (10) Hedge et al, 1973; (11) Hedge and Peterman, 1970; 
(12) Hedge et al, 1972; (13) Hart, 1973; (14) Subbarao and Hedge, 1973; (15) Subbarao, 
1972; (16) Hart et al, 1972; (17) Hedge and Peterman, 1970; (18) Hart, 1975. 



erupted material is representative of the 
source material in the mantle. This con- 
dition is met if there is isotopic equilib- 
rium between the melt and the solid 
residue in the source region so that the 
isotopic ratios either remain constant or 
change in such a manner that the effect 
can be calculated (e.g., from the known 
age of the rock) or cancelled by simple 
normalization for fractionation. Those 
who have challenged this assumption 
prefer to make a different assumption, 
namely that the source mantle is com- 
positionally uniform. O'Hara (1973, 
1975) proposed that the different mag- 
mas on Iceland and the Reykjanes Ridge 
"were derived from a homogeneous 
source, [and that] the differences could 
arise from fractional crystallization dur- 
ing the ascent of the lava." He fur- 
ther suggested that "isotope ratios had 
changed during the manifest fractional 
crystallization." Flower et al. (1975) 
"are reluctant as yet to accept the two- 
mantle-source model because it is essen- 
tially defeatist," and attempt to show 
that 87 Sr/ 86 Sr variation in the basalts 
"is not irreconcilable with a single man- 
tle composition." They do this by postu- 
lating that phlogopite, because of its 
high Rb/Sr ratio, is a source of unequili- 
brated radiogenic 87 Sr which is added to 
the melt only as the phlogopite itself en- 
ters the melt phase. O'Nions and Pank- 
hurst (1974) also find an inhomogeneous 
mantle more difficult to accept than dis- 
equilibrium melting and propose specific 
models of disequilibrium melting to ac- 
count for the observed isotopic varia- 
tions. Earlier suggestions of disequilib- 
rium melting were made by Gast et al. 
(1964) and by Peterman et al. (1970), 
though only in reference to partial fusion 
of a xenolith caught in the host basalt 
immediately prior to eruption, and by 
Harris et al. (1972), who used the evi- 
dence of isotopic disequilibrium within 
xenoliths to infer that such disequilib- 
rium might persist during partial melt- 
ing and basalt genesis. Disequilibrium 
within xenoliths or between xenolith and 

host basalt appears to be common (All- 
sopp et al., 1969; Compston and Lover- 
ing, 1969; Cooper and Green, 1969; 
Dasch and Green, 1975; Hutchison and 
Dawson, 1970; Kudo et al, 1972; Laugh- 
lin and Brookins, 1971; Leggo and 
Hutchison, 1968; Lovering and Tatsu- 
moto, 1968; Manton and Tatsumoto, 
1971; Morioka and Kigoshi, 1975; 
O'Neil et al, 1970; Paul, 1971; Peter- 
man et al., 1970; Stuckless and Ericksen, 
1975; Stueber and Ikramuddin, 1974; 
Stueber and Murthy, 1966; Stull and 
McMillan, 1973; Zartman and Tera, 
1973). Because the original relationship 
of these xenoliths to the host basalt is 
not known, extrapolation from the xeno- 
lith data to the behavior of the actual 
melting region remains dubious. 

In this controversy, relatively little 
attention has been given to the processes 
by which the mantle might be homog- 
enized. It is obvious that the process of 
withdrawing a partial melt leaves a solid 
residue that is depleted in some of its 
components. Because of this, the mantle 
regions that are the source of volcanism 
must become chemically different from 
adjacent regions. The question then 
arises, Is there a plausible mechanism by 
which these regions are rehomogenized? 
Some authors deny this (Dickinson and 
Luth, 1971; Ringwood, 1974) because 
they view the residue as a refractory 
solid that cannot be equilibrated with 
undepleted mantle because solid-state 
diffusion in silicates is known to be very 
slow. Others (Armstrong, 1968; O'Nions 
and Pankhurst, 1974) assume that man- 
tle convection combined with diffusion 
will rehomogenize an inhomogeneous 
mantle. The new diffusion data in melts 
given in the companion report by Hof- 
mann (this Report) bear on this ques- 

Most of the participants of this dis- 
cussion (O'Hara is one exception) accept 
the conclusion that the variability in 
isotopic composition of fresh volcanic 
rocks reflects isotopic inhomogeneities 
of the source. The issue is the length (or 



volume) scale of these inhomogeneities. 
Two extreme cases may be distinguished: 

(1) a mantle that is uniform on a re- 
gional scale (when comparing samples 
that are representative of 10 to 1000 
km 3 volumes) but at the same time is 
not in chemical and isotopic equilibrium 
on a local (centimeter or less) scale; 

(2) a mantle that is locally equilibrated 
but is inhomogeneous on a regional scale. 
Case 1 would be produced by very effi- 
cient mechanical stirring and extremely 
inefficient (or absent) chemical diffusion. 
Case 2 would be produced by inefficient 
mechanical stirring combined with local 
diffusion or recrystallization processes 
that lead to local chemical and isotopic 
equilibrium between individual mineral 
grains or, more importantly in the pres- 
ent context, between the melt phase and 
residual crystals. 

In this report we argue the merits of 
case 2 for a partially molten mantle. For 
this purpose, we examine the regional 
pattern of isotopic compositions in 
oceanic rocks, the constraints set by 
isotopic model ages, and the shortcom- 
ings of published models that are based 
on the assumption of case 1. In addition 
we apply available kinetic data to esti- 
mate the length scale of equilibration in 
a partially molten mantle. We are led to 
the conclusion that the isotopic composi- 
tion of oceanic volcanic rocks is indeed 
representative of the source and that the 
mantle is inhomogeneous (on a regional 
scale) both horizontally and vertically. 

Evidence Based on Analytical Data 
from Volcanic Rocks 

The idea of a chemically heterogeneous 
mantle received its first strong verifica- 
tion from the Sr and Pb isotope work of 
Faure and Hurley (1963), Hedge and 
Walthall (1963), and Cast, Tilton, and 
Hedge (1964). Recent high precision 
isotopic work has not markedly changed 
the picture, though it has added consid- 
erable detail to earlier findings. We will 
restrict our discussion to oceanic vol- 
canics, as there is less possibility that 

crustal contamination is involved in 
these cases. There is no convincing evi- 
dence of crustal contamination in any of 
the oceanic volcanic isotopic data and, 
as will be pointed out later, the lead 
isotope data appear to preclude any sig- 
nificant contamination. Other evidence, 
such as the striking uniformity of Sr 
isotope ratios for recent tholeiites from 
the whole length of the active volcanic 
zone in Iceland (Hart et al., 1973), also 
suggests involvement of crustal contami- 
nants is minimal or absent. 

The lead isotope data of oceanic vol- 
canic rocks have been recently sum- 
marized by Sun and Hanson (1975). In 
addition, we have collected most of the 
recent high precision Sr isotope data 
(that is, data published since 1970) for 
oceanic islands and mid-ocean ridges, 
and these data are summarized in Fig. 
47. This figure simply demonstrates 
again the well-known variability of Sr 
isotope ratios in oceanic rocks and em- 
phasizes the clear-cut difference between 
mid-ocean ridge basalts (MORB) and 
island basalts. The demarcation between 
ridge basalts and island basalts occurs 
at a 87 Sr/ 8(5 Sr value of about 0.7030. 
Very few island basalts have values be- 
low 0.7030; two exceptions are St. Helena 
and Ascension Islands, which are sig- 
nificantly below 0.7030. There are a 
number of ridge basalts with values 
higher than 0.7030, but these are invari- 
ably associated either with seawater 
alteration or proximity to hot spots or 
plumes. For example, some of the high 
values reported by Hedge and Peterman 
(1970) are probably altered, as judged 
by petrographic descriptions and Rb and 
Cs contents reported in Kay, Hubbard, 
and Gast (1970). Other high values oc- 
cur in samples probably affected by 
nearby hot spots; for example, the sam- 
ples from 45°N on the mid-Atlantic 
ridge appear to be related to the Azores 
hot spot (White, Hart, and Schilling, 
1975), while samples from the mid- 
Indian Ocean ridge (Subbarao and 
Hedge, 1973) are probably related to 


either the Reunion hot spot or to the One aspect of the isotopic data that 
triple junction on the mid-Indian Ocean bears on the problem of scale of the 
ridge. To our knowledge, there is no mantle heterogeneities is the marked 
well-documented Sr isotope ratio on uniformity of Sr isotopic ratios for 
fresh MORB which is higher than 0.7030, basalts from a given geographic area de- 
when only ridge segments well removed spite large variations in the other chemi- 
from hot spots are considered. cal constituents of the basalt. For ex- 
This, then, is the type of data which ample, tholeiites from the whole length 
demonstrates isotopic heterogeneity from of the neo-volcanic zone on Iceland are 
the sub-oceanic mantle. By itself, it does isotopically uniform (within the limits 
not settle the question of whether that of measurement) despite considerable 
heterogeneity is on a mineralogical (= variation in other trace element concen- 
local) scale or on a larger, regional scale, trations (Hart et al., 1973; Hart and 
If the isotopic differences between ridge Schilling, 1973). With rather limited 
basalts reflect regional mantle hetero- sampling, we have already demonstrated 
geneities, then it would appear difficult potassium contents ranging from 150 
to keep these sources distinct in a con- ppm to 2000 ppm (and an even larger 
vecting mantle, as islands may * occur range is shown by the data of Sigvalda- 
very close to ridges. This problem was son, 1974), with like variations in other 
posed by Hart (1971) and has recently LIL (large ion lithophile) element con- 
been the main consideration leading centrations such as Rb, Cs, and Ba. This 
O'Nions and Pankhurst (1974) to con- phenomenon can be extended to some of 
sider disequilibrium melting models as the alkali basalts on Iceland, which have 
an alternative to the generally accepted Sr isotope ratios similar to the tholeiites 
idea of a regionally heterogeneous man- despite their markedly different chemi- 
tle. Another, and in our opinion more cal character. If a small-scale disequi- 
plausible, solution to this problem is librium model (corresponding to case 1 
given by a vertically stratified mantle, in the "Introduction") is invoked to ex- 
A specific model (see for example, Schill- plain the Sr isotopic data from Iceland, 
ing, 1973, and Sun and Hanson, 1975) this model must at the same time ex- 
of such a mantle may consist of a con- plain how such diverse chemical charac- 
vecting upper mantle that is chemically teristics are obtained from a chemically 
depleted by previous melting and a more homogeneous mantle. While O'Nions 
"juvenile" lower mantle, which may or and Pankhurst (1974) suggest that the 
may not convect separately. Mid-ocean Sr isotopic characteristics of oceanic ba- 
ridge basalts with their low 87 Sr/ 86 Sr salts may be explained by disequilibrium 
ratios and depleted alkalies and other melting models, they do not show how 
trace elements are derived from the up- such a model can be made compatible 
per mantle; oceanic island basalts with with the isotopic uniformity coupled 
their higher 87 Sr/ 86 Sr ratios and their with chemical diversity of basalts from 
less depleted trace element character are Iceland. A similar situation exists in the 
derived from plumes or isolated diapiric region of the Azores hot spot (White, 
intrusions that rise from the lower man- Hart and Schilling, 1975), where tho- 
tle with relatively little chemical inter- leiites dredged from the Azores platform 
action with the upper mantle. A model have Sr isotope ratios identical to those 
like this, however, requires many plumes of the alkali basalts from the nearby 
or diapirs, as all oceanic islands and islands, despite rather large differences 
probably many seamounts will require in chemical character, 
derivation from a mantle source that is A related phenomenon is that most 
chemically distinct from the spreading islands show considerable uniformity of 
ridge mantle. Sr isotope abundances, whereas separate 



islands will have distinctly different Sr 
ratios. This can be seen for the various 
islands of the Azores group (White, 
Hart, and Schilling, 1975), as well as 
other islands in the Atlantic (O'Nions 
and Pankhurst, 1974). In other words, if 
disequilibrium melting of a regionally 
homogeneous mantle is envisaged, it has 
to be carried out in such a way that Sr 
isotope ratios show inter-island but not 
intra-island variations. It is conceivable 
and has been vigorously argued by 
O'Hara (1973, 1975) that the intra- 
island chemical differences are the result 
of different degrees of fractional crystal- 
lization. In this case, the isotopic com- 
position of the residual melts could re- 
main constant despite their chemical 
diversity. Nevertheless, this does not 
solve the basic problem of obtaining the 
isotopic uniformity in the first place, 
namely at the melting stage through a 
disequilibrium process. If the melts are 
to be isotopically uniform but different 
from this source, the extent of reaction 
between solid and (disequilibrium) melt 
fractions must be remarkably uniform 
for all batches of magma within each 
island. At the very least, both the rate 
of melt production and the degree of 
partial melting would have to be uni- 

In disequilibrium melting models, 
phlogopite generally plays a key role 
(O'Nions and Pankhurst, 1974; Flower, 
Schmincke, and Thompson, 1975). The 
crucial question here is of course whether 
phlogopite can remain out of isotopic 
equilibrium for times of the order of 10 7 
to 10° years, and whether it can remain 
as a residual phase during any signifi- 
cant degree of melting. We will address 
the first part of this question below; the 
second part requires experimental veri- 
fication. Here, we only point out that 
the amount of phlogopite in the mantle 
should not be treated as an uncon- 
strained variable. For example, O'Nions 
and Pankhurst (1974) consider 5% 
phlogopite in their mantle model. Is this 
reasonable? Phlogopite contains about 

8% potassium, and the overall K con- 
tent of the postulated mantle rocks is 
directly related to phlogopite content. 
A peridotite containing 5% phlogopite 
will have a K content of 4000 ppm. This 
is probably 10 times the K content of al- 
most any mantle-derived ultramafic rock 
yet analyzed, and 4 to 5 times the K 
content of chondrites. Ever since Gast's 
(1960) classic paper it has been accepted 
that the earth's crust and upper mantle 
are depleted in alkalies relative to 
chondrites. A K content of 4000 ppm 
implies heat production of at least 4-5 
times the chondritic value (or more, if a 
terrestrial K/U ratio is assumed) ; a 
model for a sub-oceanic mantle with this 
level of heat production would be seri- 
ously out of step with estimates based 
on observed heat flow. The arguments 
of Gast (1960) and subsequent workers 
need not be restated here, and we add 
only that unless entirely new evidence is 
introduced, the chondritic K value of 
~0.08% must be considered an upper 
limit for the terrestrial K abundance. 
Consequently, if we consider a mantle 
with a maximum of 1% phlogopite as 
the source of the melts, a very small de- 
gree of partial melting will consume all 
the available phlogopite, unless the 
phlogopite is far more refractory than is 
indicated by experimental studies on 
phlogopite stability (Yoder and Kushiro, 
1969; Modreski and Boettcher, 1973; 
Forbes and Flower, 1974). 

Evidence Derived from Time Constraints 

It is obvious that time considerations 
are important in this discussion, because 
isotopic and trace element equilibrium 
or disequilibrium will be strongly de- 
pendent on kinetic factors. Any serious 
chemical model for the mantle must be 
consistent with the fact that the ob- 
served heterogeneities have persisted for 
times of the order of 1-3 billion years. 
The most persuasive evidence for this is 
the general alignment of volcanic lead 
isotope data along a secondary isochron 
of about 2 billion-year slope. This argu- 



ment was originally presented by Gast, 
Tilton, and Hedge (1964) and Tatsu- 
moto (1966a, 19666). Since that time, 
numerous papers have added Pb isotope 
data from a variety of volcanic environ- 
ments, and have gradually improved the 
precision of the isotopic data (see, for 
example, Sun and Hanson, 1975). Data 
from oceanic islands and oceanic ridges, 
when plotted on a 2(Hi Pb/ 204 Pb versus 
207 Pb/ 204 Pb diagram, fall along a linear 
trend which, interpreted as a secondary 
isochron, has an age of ~2.0 billion 
years. These lead data require deriva- 
tion of lead from sources which have 
maintained separate and distinct U/Pb 
systems for this length of time. Unless 
one of these sources is considered to be 
crustal (i.e., sediment, seawater, etc.), 
the source of the Pb isotope heterogenei- 
ties must reside in the mantle. One argu- 
ment against sediment involvement is 
the near absence of sediments in the 
vicinity of the mid-ocean ridge. Direct 
involvement of seawater lead is also un- 
likely, because the concentration of Pb 
in seawater is on the order of 30,000 
times lower than in basalt, and inter- 
action with the very large volumes of 
seawater required would change the 
oxygen isotopic composition of the ba- 
salt profoundly. Another argument can 
be made from recent comparisons be- 
tween volcanic leads and ocean sediment 
leads (Sinha and Hart, 1972; Church 
and Tilton, 1973; Meijer, 1975). In gen- 
eral, oceanic sediment leads (and pre- 
sumably seawater Pb as well, Chow and 
Patterson, 1962) lie along a linear trend 
which is subparallel to the oceanic vol- 
canic lead line, but which has signifi- 
cantly higher values of 207 Pb/ 204 Pb. 
Thus the crustal lead available for con- 
tamination of ocean volcanic rocks does 
not fall either along or on an end- 
member to the oceanic trend, but above 
it (on a 207 Pb/ 204 Pb- 206 Pb/ 204 Pb plot) 
and subparallel to it. 

Because only one radioactive parent 
isotope is involved in the Sr isotope sys- 
tem, one cannot make the same argu- 

ment as with the lead. However, as 
shown by Sun and Hanson (1975), the 
oceanic volcanic Sr data fall approxi- 
mately along a 2.0 b.y. isochron when 
plotted on a Rb-Sr isochron diagram. 
This age is meaningful only if the Rb/Sr 
ratio of the volcanic liquid is the same 
as that of the source. However, for the 
age of the mantle heterogeneities to be 
substantially less than this 2.0 b.y. value, 
it is necessary to postulate a very sys- 
tematic type of disequilibrium melting 
which can produce liquids with Rb/Sr 
ratios that are systematically lower than 
in the source so that the approximate 
isochron relationship is preserved. Equi- 
librium melting involving residual phlog- 
opite is not an effective way of pro- 
ducing liquids with Rb/Sr ratios signifi- 
cantly lower than the source (retention 
of 2°/o phlogopite after melting of a 
source containing an original 5% phlog- 
opite will decrease the Rb/Sr ratio of 
the liquid by only 20% relative to the 
starting material). 

Evidence for mantle heterogeneity 
(on some scale) exists not only from Pb 
and Sr isotope evidence, but also from 
trace element ratios and abundances. 
While one cannot prove that the trace 
element heterogeneities have also existed 
for a long time period, as one can for the 
lead isotope data, it is probably reason- 
able to consider all of the heterogeneities 
as being related to the same processes. 
After all, the Pb and Sr isotopic varia- 
tions are generated by variations in 
U/Pb and Rb/Sr, and these four trace 
elements are similar in general behavior 
to the many other trace elements for 
which mantle heterogeneity is postulated 
(K, Cs, Ba, REE, etc.). Thus we feel 
that any discussion of mechanisms for 
generating a particular trace element 
pattern in volcanic rocks should be com- 
patible with the constraints imposed by 
time. For example, O'Nions and Pank- 
hurst (1974) consider a model in which 
disequilibrium melting of phlogopite 
peridotite occurs, with the disequilibrium 
having persisted for only 20 m.y. Obvi- 



ously, it is much easier to conceive of 
inter-mineral disequilibrium lasting for 
20 m.y. than for 2000 m.y. We argue 
that models such as proposed by O'Nions 
and Pankhurst be considered only if 
they are consistent with disequilibrium 
on a time scale of approximately 2000 

O'Hara's (1973, 1975) explanation for 
the trace element and isotopic variations 
between tholeiites from Iceland and 
those from the Reykjanes Ridge requires 
special discussion, because he assumed 
a uniform mantle source for the entire 
region and relied entirely on fractional 
crystallization to explain all the chemi- 
cal variation. Hart et al. (1973) ex- 
plained the Sr isotope data from these 
basalts as due to large-scale heteroge- 
neities in the mantle. O'Hara (1973) 
first discounted the isotopic evidence by 
alluding to the implicit "assumption that 
these isotopes cannot be fractionated." 
Actually, the question of whether Sr iso- 
topes can be fractionated in nature or 
not is irrelevant, because all Sr isotope 
analyses are routinely corrected for all 
isotopic fractionation effects, be they 
natural or laboratory induced. O'Hara 
(1975) clarified his earlier statement, 
stating the process was one involving a 
change of isotope ratio "during manifest 
fractional crystallization." Presumably 
this refers not to isotope fractionation but 
to generation of radiogenic 87 Sr during 
histories in variable Rb/Sr environ- 
ments. To produce the observed differ- 
ence in Sr isotope ratios of the Iceland- 
Reykjanes Ridge basalts (0.7027 versus 
0.7030) would require aging for 15 m.y. 
in an environment having a Rb/Sr ratio 
of 0.5. The oldest rocks in Iceland are 
about 15 m.y. old (O'Nions and Pank- 
hurst, 1973), so this is a reasonable up- 
per limit for the age of a Rb/Sr subsys- 
tem. However, the Rb/Sr ratio of 0.5 is 
at least 10 times the highest ratio ob- 
served in any of the basalts from Ice- 
land. O'Hara's model will only work if 
one assumes that the Rb source which 
supported the growth of 87 Sr in Icelandic 

basalts is now hidden. O'Hara's model is 
even more difficult to defend when the 
Pb isotope data are considered as well. 
Sun and Jahn (1975) have shown that 
the Sr isotope variation between Iceland 
and the Reykjanes Ridge is paralleled 
by a variation in Pb isotope abundances. 
The Pb isotope data form a linear trend 
similar to that commented on above for 
oceanic volcanics in general, with an 
inferred age of about 1.5 b.y. Obviously, 
short-term fractionation models such as 
O'Hara's do not explain the recent Ice- 
land isotopic data nor the more general 
body of geochemical data (of which Ice- 
land is only a specific case) that has 
been available for over ten years. 

Evidence for Mineral Disequilibrium 

Perhaps the closest analog to possible 
mantle material is found in the ultra- 
mafic nodules which are erupted in alkali 
basalts and other undersaturated basalt 
types. The large majority of these nod- 
ules appear isotopically different from 
their host basalts (these findings are 
based on at least 19 different publica- 
tions covering about 100 separate nod- 
ules). Most of these studies involved Sr 
isotope data ; however, lead isotope stud- 
ies on nodule suites (e.g., Zartman and 
Tera, 1973) also indicate disequilibrium 
between nodules and host basalt. Be- 
cause of the absence of any systematic 
relationship between nodules and host, 
the authors of these studies have gen- 
erally concluded that the nodules are not 
related genetically to their host basalt 
but are probably accidental inclusions. 

In addition to disequilibrium between 
nodule and host basalt, 10 of the studies 
have analyzed the constituent minerals 
of the nodules (usually olivine, orthopy- 
roxene, and clinopyroxene) and in nearly 
all cases disequilibrium was also found 
to exist between the minerals of a given 
nodule. Again, most of these studies re- 
lied on Sr isotope data; however, similar 
results were reported for Pb isotope 
abundances by Zartman and Tera (1973) 
and by Morioki and Kigoshi (1975). 



Generally, the olivine and orthopyroxene 
have higher 87 Sr/ 86 Sr ratios than coexist- 
ing clinopyroxene ; in some cases the dif- 
ferences are very large. For example, a 
nodule from Victoria, Australia (Dasch 
and Green, 1975), contained olivine with 
87 Sr/ 8(5 Sr = 0.713, orthopyroxene = 
0.7060, clinopyroxene = 0.7029, total 
bulk nodule = 0.7040, and host basalt 
= 0.7045. While this particular case 
nicely illustrates the Sr isotope disequi- 
librium that is frequently found between 
minerals in nodules, it also illustrates 
some of the unanswered problems. A 
mineralogical mode was reported for this 
nodule, and when the individual mineral 
analyses are used to reconstruct the 
whole rock, the Sr and Rb concentra- 
tions are lower by a factor of about 3 
than the directly analyzed whole rock 
nodule, while the resulting 87 Sr/ 8G Sr 
ratio (0.7055) is higher than the bulk 
nodule. This implies that some phase or 
material is present in the nodule which 
was not analyzed but which has rela- 
tively high Rb and Sr contents, and an 
87/86 ratio lower than 0.7040. While 
this question does not directly affect the 
concept of mineral disequilibrium in 
nodules, it does show that nodules are 
complex systems and allows room for 
speculation about possible contaminat- 
ing effects of such material on separated 
minerals. It should be noted that al- 
though most nodules that have been 
studied come from basalts erupted 
through continental terrains, disequilib- 
rium has been proven in at least one case 
for an oceanic site, Hawaii (Zartman 
and Tera, 1973; Stueber and Ikramud- 
din, 1974). 

Inter-mineral disequilibrium in ultra- 
mafic nodules does not demonstrate, 
however, that the mantle in magma 
source regions will show similar dis- 
equilibrium. Because most of these 
nodules are not cogenetically related to 
their host basalt, and because their con- 
stituent minerals are not in equilibrium, 
little can be said about the depth from 
which they come, or whether they do in 

fact come from a mantle region which is 
presently acting as a source for magmas. 
They could come from the uppermost 
mantle levels of the lithosphere where 
temperatures have been relatively low 
(500°-600°C) for long periods of time 
(subcontinental lithosphere is essentially 
protected against high temperatures and 
reworking by virtue of the overlying 
continent). Even in Hawaii (where the 
observed isotopic disequilibrium is less 
than most of the continental occur- 
rences), the lithosphere is probably 10 8 
years, or older. One obvious test of the 
possibility that most nodules have not 
been at magmatic temperatures for a 
long time would be to analyze nodules 
from an area of very young lithosphere 
— for example, an island which is very 
close to a spreading ridge. If disequilib- 
rium exists in such nodules, one can be 
fairly sure of their recent high-tempera- 
ture history. Nevertheless, even at tem- 
peratures just below the solidus, the 
equilibration rate is expected to be much 
lower than in a partially molten state. 
This point is explained more fully below. 

Application of Kinetic Data 

Although reliable experimental stud- 
ies on the kinetics of reactions and dif- 
fusion are scarce, the available data all 
appear to point to local equilibrium, at 
least if grain diameters are on the order 
of 1 cm or less and if the time available 
for melting is on the order of a few 
thousand years or more. Several lines of 
evidence lead to this conclusion: 

1. Melting experiments conducted with 
natural or artificial peridotite systems 
generally attain equilibrium within hours 
or days (e.g., Mysen and Boettcher, 
1975). The melting process forces all 
major residual solid phases to change 
composition. Thus, zoned crystals can 
be observed when equilibrium has not 
been attained. The slowest possible proc- 
ess for this change in composition is 
probably volume diffusion. The actual 
process in the presence of a melt is likely 
to be dissolution into the melt and re- 



precipitation of the melt-equilibrated 
solid, whenever the coherency stress with 
the crystal between initial and equili- 
brated composition is sufficiently large 
(Petrovic, 1973). If the rate-limiting 
step is diffusion, the equilibration time 
can be extrapolated to coarser grained 
material. Typical grain diameters used 
in electron microprobe examination of 
experimental materials are about 50 /mi. 
An extrapolation to 1 cm requires a 
time factor of about 40,000, so that the 
time for equilibration is on the order of 
a hundred years. 

2. Direct determinations of diffusion 
coefficients are available for some of the 
phases of interest. These are shown in 
Fig. 48. It appears from these results 
that at 1000°C, the probable lower tem- 
perature limit for melting in the mantle, 
D = 10~ 13 cm 2 sec -1 may be taken as 
a lower limit for cation diffusion in man- 
tle minerals. This means that a sphere 
of 1 cm diameter will equilibrate with a 
well-mixed melt phase in less than 4 X 
10 4 years. The word "equilibrate" is 
used here in the sense that the actual 
mass transfer between crystal and en- 
vironment is 99% of that required for 
complete equilibration. 

3. Sr diffusion in phlogopite is critical 
to several of the nonequilibrium-melting 
models proposed by O'Nions and Pank- 
hurst (1974) and Flower et al (1975) be- 
cause, if present, phlogopite is expected 
to be the major contributor of radio- 
genic strontium in the mantle. If some 
of this 87 Sr can be retained in the phlog- 
opite before and during partial melting, 
or if it does not equilibrate with the non- 
radiogenic strontium of another phase 
such as clinopyroxene, the 87 Sr/ 86 Sr ratio 
of the melt may either increase or de- 
crease as the degree of melting increases, 
depending on the particular process of 
nonequilibrium melting. To the best of 
our knowledge, the radiogenic argon and 
strontium of all mica minerals is either 
lost or equilibrated with the environ- 
ment during any high-grade metamorphic 
event. This is borne out by radiometric 


I0" 9 h 

M 10 






1600 1400 1200 1000 800 

" \ ' V 'i ' i ' 

7 9 

!0 4 /T(°K) 

Fig. 48. Arrhenius plot (log D versus 1/T) 
for diffusion in silicates relevant to mantle 
rocks. Solid lines are experimentally deter- 
mined; dashed lines are inferred or extrapo- 
lated. A, Ar in phlogopite (Giletti, 1974b) ; B, 
Sr in phlogopite estimated from line A and 
from data by Hart (1964); C, Fe-Mg inter- 
diffusion in olivine (Fo 87 ; /o 2 = 10~ 12 atm) 
from Buening and Busek (1973); D, FeMg 
interdiffusion in olivine (Fo 0O ) from Misener 
(1974); E, U in diopside (Seitz, 1973); F, Al 
in diopside (Seitz, 1973) ; G, H, and /, oxygen 
in enstatite, diopside, and forsterite, respec- 
tively (Muehlenbachs, 1974) ; J, Ca in basalt 
(Hofmann, this Report). 

age determinations by numerous authors, 
and it is true even for short-term heating 
during contact metamorphism (e.g., 
Hart, 1964; Hanson and Gast, 1967). 
From this evidence alone it appears ex- 
tremely unlikely that phlogopite can 
retain measurable amounts of radiogenic 
strontium at temperatures above 1000°C. 
To obtain an approximate time for Sr 
equilibrium between phlogopite and its 


environment, we extrapolate the diffu- tors: (1) the extremely low concentra- 
sion measurement of Giletti (19746) on tion of Sr in olivine and orthopyroxene, 
argon in phlogopite to higher tempera- the two major phases present; and (2) 
tures and estimate that the diffusion the potentially much greater distance 
coefficient for Sr is no more than a factor over which strontium must diffuse 
of 10 lower. This estimate is derived through solid material. To illustrate the 
from the well-determined relative be- first factor we imagine a clinopyroxene 
havior of Sr and Ar in biotite (Hart, grain separated from a phlogopite grain 
1964; Hanson and Gast, 1967). Giletti's by a barrier of olivine. If the intrinsic 
results are used instead of other pub- mobility of Sr in all three phases is the 
lished data because his measurements same but the concentration in the olivine 
are the only ones in which the mica was is lower by a factor of 500 (see parti- 
stable under the experimental conditions, tioning data by Hart and Brooks, 1974) , 
The estimated diffusion coefficients for the rate of exchange or flux through the 
Sr in phlogopite are shown in Fig. 48. olivine barrier is reduced by a factor 
The resulting maximum time for Sr of 500. The second factor is important 
equilibration between phlogopite and because diffusion times vary with the 
melt at 1000°C is about 1200 years for square of the diffusion distance. Wilshire 
a sphere or 1900 years for a cylinder of and Jackson (1975) show examples of 
1 cm diameter. compositionally inhomogeneous xenoliths 

Studies of trace element partitioning that would require diffusion over dis- 

(e.g., Philpotts and Schnetzler, 1970; tances of 10 cm or more for homogeniza- 

Shimizu, 1974; Hart and Brooks, 1974) tion to occur. Once a sufficient amount 

have shown that clinopyroxene is prob- of melt is present to wet all the grain 

ably the major Sr-bearing phase in the boundaries, then barriers to internal 

upper mantle. Slow diffusion in clino- equilibration are removed since the melt 

pyroxene might therefore inhibit overall phase acts as a short circuit between the 

equilibration even if diffusion in phlogo- solids because diffusion is several orders 

pite and other phases is rapid. Argon of magnitude faster in the silicate melt 

retention in pyroxenes is known to be than in the solids (see Fig. 48). Conse- 

higher than in micas, and it is conceiv- quently, the presence of a melt phase 

able that the strontium in clinopyroxene may be necessary to the establishment 

does not equilibrate with the melt; we of local equilibrium in the mantle, 

are therefore planning to determine Sr This leaves the question of large-scale 

diffusion in clinopyroxene experimentally, rehomogenization in the mantle. O'Nions 

It should be remembered, however, that and Pankhurst (1974) argue the case for 

extraordinarily small diffusion coefficients local disequilibrium for short-lived proc- 

(D < 10 -15 cm 2 sec -1 ) are required for esses (e.g., accumulation of radiogenic 

disequilibrium to persist between liquid Sr in phlogopite equivalent to an age 

and a coarse-grained solid (grain diam- of 20 m.y.). At the same time they be- 

eter > 1 cm) for a time span of one lieve that, on a regional (kilometer) 

million years or more. scale, the mantle will be homogenized 

If one accepts a diffusion coefficient by the combined effects of diffusion and 

of 10 -13 cm 2 sec -1 at 1000°C, the ap- convection. The times involved for this 

parent disequilibrium in mantle xenoliths process are presumably somewhere be- 

is not altogether surprising, if one as- tween 10 8 and 5 X 10 9 years. As pointed 

sumes that the xenolith existed in a fluid- out above, diffusion distances vary with 

free environment before being caught up the square root of time. Using again D 

by the rapidly intruding melt. In an all- = 10 -13 cm 2 sec -1 , a characteristic dis- 

solid peridotite, Sr isotopic equilibration tance of Dt = ~60 cm is obtained for 

between phases is inhibited by two fac- a diffusion time of 10 9 years. Larger 


scale homogenization is therefore out of rock are sampled and homogenized dur- 
the question for an all-solid mantle. This ing separation of the melt from its solid 
statement must be qualified by adding residue and intrusion to the surface. (3) 
that we assume that mantle convection The depleted nature of mid-ocean ridge 
without diffusion is not effective in pro- material relative to ocean-island material 
ducing large-scale homogeneous mantle requires a vertically stratified mantle, 
corresponding to case 1 discussed in the If the two types of material happen to 
introduction. We infer this qualitatively extrude in close mutual proximity, they 
from the present-day surface kinematics will nevertheless retain their isotopic iden- 
of plate movement, which point to large tities. (4) The processes of subduction 
convection cells (on the order of 1000 and production of oceanic islands give 
km) and slow turnover time (~10 8 rise to a heterogeneous mantle. Reho- 
years) . Even models that consider two mogenization by convection and diffusion 
scales of convection (Richter and Par- is unlikely because diffusion is too slow, 
sons, 1975) envision small rolls greater and models that require a homogeneous 
than 100 km in diameter. In the presence mantle appear to be unrealistic. (5) Once 
of a melt, the diffusion coefficients and, a heterogeneous mantle is accepted there 
therefore, the distances are substantially is no need to postulate local disequi- 
higher, but even in this case, regional librium during melting or isotopic 
homogenization is achieved only by changes (as proposed by O'Hara, 1973, 
rather vigorous stirring. The character- 1975) during subsequent fractional crys- 
istic distance for a partially molten tallization. (6) The geophysical conse- 
mantle, calculated using D = 10 -6 cm 2 quences of a stratified mantle need fur- 
sec -1 as an upper limit for the diffusivity ther exploration. The stratification may 
(see companion report by Hofmann), is be stabilized by the relatively greater 
\/Dt = 1 km. It is therefore difficult, density of "fertile" mantle as compared 
for example, to see how a subducted with "depleted" mantle (O'Hara, 1975), 
oceanic crust of 5 km thickness can ever and the two or more layers may form 
be equilibrated with the surrounding separate convection systems. Intrusion 
mantle as proposed by Armstrong (1968). from the lower, fertile layer through the 
Extensive mixing appears to be plausible depleted layer would then require a 
only in environments where very large small amount of partial melting at the 
shear strains are encountered, for exam- top of the fertile layer or an adiabatic 
pie, within the diapiric intrusion of par- rise of plumes from even greater depths, 
tially molten material. On the other Again, such a plume would be partially 
hand, little or no isotopic exchange is molten by the time it encounters the 
expected to occur between separate dia- depleted upper mantle, 
pirs, even if they are in close proximity. We agree with O'Nions and Pankhurst 
We conclude that the observed isotopic (1974) that the isotopic data do not by 
patterns of oceanic volcanic rocks are themselves prove or disprove the exist- 
altogether unsurprising: (1) Individual ence of plumes; but if our conclusions 
oceanic islands frequently show constant regarding local equilibrium are confirmed 
isotopic composition regardless of rock by further experimental work on kinetics, 
type (Hart et al., 1973; O'Nions and the isotopic data can serve as a powerful 
Pankhurst, 1974) because of mixing dur- constraint on the choice of convection 
ing diapiric rise of the source material, pattern in the mantle. 
Each diapir samples a different portion 

of a heterogeneous mantle. (2) Mid- References 

ocean ridge material is more uniform on Allsopp, H. L., L. O. Nicolaysen, and P. 

a larger scale than oceanic islands be- Hahn-Weinheimer, Rb/K ratios and 

cause much larger volumes of source Sr-isotopic compositions of minerals 



in eclogitic and peridotitic rocks, Earth 
Planet. Sci. Lett., 5, 231-244, 1969. 

Armstrong, R. L., A model for Sr and Pb 
isotope evolution in a dynamic earth, 
Rev. Geophys., 6, 175-199, 1968. 

Buening, D. K., and P. R. Buseck, Fe- 
Mg lattice diffusion in olivine, J. Geo- 
phys. Res., 78, 6852-6862, 1973. 

Chow, T. J., and C. Patterson, The oc- 
currence and significance of lead iso- 
topes in pelagic sediments, Geochim. 
Cosmochim. Acta, 26, 263-308, 1962. 

Church, S. E., and G. R. Tilton, Lead 
and strontium isotopic studies in the 
Cascade Mountains; bearing on ande- 
site genesis, Geol. Soc. Am. Bull., 84, 
431-454, 1973. 

Compston, W., and J. F. Lovering, The 
strontium isotopic geochemistry of 
granulitic and eclogitic inclusions from 
the basic pipes at Delegate, Eastern 
Australia, Geochim. Cosmochim. Acta, 
33, 691-699, 1969. 

Cooper, J. A., and D. H. Green, Lead 
isotope measurements on lherzolite in- 
clusions and host basanites from west- 
ern Victoria, Australia, Earth Planet. 
Sci. Lett., 6, 69-76, 1969. 

Dasch, E. J., and D. H. Green, Strontium 
isotope geochemistry of lherzolite in- 
clusions and host basaltic rocks, Vic- 
toria, Australia, Am. J. Sci., 275, 461- 
469, 1975. 

Dickinson, W. R., and W. C. Luth, A 
model for plate tectonic evolution of 
mantle layers, Science, 174, 400-404, 

Faure, G., and P. M. Hurley, The iso- 
topic composition of strontium in oce- 
anic and continental basalts: applica- 
tion to the origin of igneous rocks, J. 
Petrol, 4, 31-50, 1963. 

Flower, M. F. J., H.-U. Schmincke, and 
R. N. Thompson, Phlogopite stability 
and the 87 Sr/ 86 Sr step in basalts along 
the Reykjanes Ridge, Nature, 254, 
404-406, 1975. 

Forbes, W. C, and M. F. J. Flower, 
Phase relations of titanphlogopite, K 2 
Mg 2 TiAl 2 Si 6 O 20 (OH) 4 : a refractory 

phase in the upper mantle? Earth 
Planet. Sci Lett., 22, 60-66, 1974. 

Gast, P. W., Limitations on the composi- 
tion of the upper mantle, J. Geophys. 
Res., 65, 1287-1297, 1960. 

Gast, P. W., G. R. Tilton, and C. Hedge, 
Isotopic composition of lead and 
strontium from Ascension and Gough 
Islands, Science, 145, 1181-1185, 1964. 

Giletti, B. J., Studies in diffusion I: 
argon in phlogopite mica, in Geochemi- 
cal Transport and Kinetics, A. W. 
Hofmann, B. J. Giletti, H. S. Yoder, 
Jr., and R. A. Yund, eds., Carnegie 
Inst. Wash. Publ. 634, PP- 107-115, 

Grant, N. K., J. L. Powell, J. V. Walther, 
and F. R. Burkholder, The isotopic 
composition of strontium in lavas from 
St. Helena, South Atlantic, submitted 
to Earth Planet. Sci. Lett., 1974. 

Hanson, G. N., and P. W. Gast, Kinetic 
studies in contact metamorphic zones, 
Geochim. Cosmochim. Acta, 31, 1119— 
1153, 1967. 

Harris, P. G., R. Hutchison, and D. K. 
Paul, Plutonic xenoliths and their re- 
lation to the upper mantle, Philos. 
Trans. R. Soc. London, Ser. A, 271, 
313-323, 1972. 

Hart, S. R., The petrology and isotopic- 
mineral age relations of a contact zone 
in the Front Range, Colorado, J. Geol., 
72, 493-525, 1964. 

Hart, S. R., K, Rb, Cs, Sr and Ba con- 
tents and Sr isotope ratios of ocean 
floor basalts, Philos. Trans. R. Soc. 
London, Ser. A, 268, 573-587, 1971. 

Hart, S. R., Submarine basalts from 
Kilauea rift, Hawaii: Nondependence 
of trace element composition on extru- 
sion depth, Earth Planet. Sci. Lett., 
20, 201-203, 1973. 

Hart, S. R., LIL geochemistry of Leg 
34 basalts, Initial Reports of the Deep 
Sea Drilling Project, 34, in press, 1975. 

Hart, S. R., and C. Brooks, Clinopyrox- 
ene-matrix partitioning of K. Rb, Cs, 
Sr and Ba, Geochim. Cosmochim. Acta, 
38, 1799-1806, 1974. 

Hart, S. R,, W. E. Glassley, and D. E. 



Karig, Basalts and sea-floor spreading 
behind the Mariana Island arc, Earth 
Planet. Sci. Lett., 15, 12-18, 1972. 

Hart, S. R., and J.-G. Schilling, The geo- 
chemistry of basalts from Iceland and 
the Reykjanes Ridge, Carnegie Inst. 
Wash. Year Book 72, p. 259, 1973. 

Hart, S. R., J.-G. Schilling, and J. L. 
Powell, Basalts from Iceland and along 
the Reykjanes Ridge: Sr isotope geo- 
chemistry, Nature, Phys. Sci., 246, 
104-107, 1973. 

Hedge, C. E., and Z. E. Peterman, The 
strontium isotopic composition of ba- 
salts from the Gordo and Juan de 
Fuca Rises, northeastern Pacific Ocean, 
Contrib. Mineral. Petrol, 27, 114-120, 

Hedge, C. E., Z. E. Peterman, and W. 
R. Dickinson, Petrogenesis of lavas 
from western Samoa, Geol. Soc. Am. 
Bull, 83, 2709-2714, 1972. 

Hedge, C. E., and F. G. Walthall, Radio- 
genic strontium-87 as an index of 
geologic processes, Science, 140, 1214- 
1217, 1963. 

Hedge, C. E., N. D. Watkins, R. A. 
Hildreth, and W. P. Doering, 87 Sr/ 86 Sr 
ratios in basalts from islands in the 
Indian Ocean, Earth Planet. Sci. Lett., 
21, 29-34, 1973. 

Hurley, P. M., H. Hughes, G. Faure, H. 
Fairbairn, and W. Pinson, Radiogenic 
strontium-87 model for continent for- 
mation, J. Geophys. Res., 67, 5315- 
5336, 1962. 

Hutchison, R., and J. B. Dawson, Rb, 
Sr and 87 Sr/ 86 Sr in ultrabasic xeno- 
liths and host-rocks, Lashaine Vol- 
cano, Tanzania, Earth Planet. Sci. 
Lett, 9, 87-92, 1970. 

Kay, R., N. J. Hubbard, and P. W. Gast, 
Chemical characteristics and origin of 
oceanic ridge volcanic rocks, J. Geo- 
phys. Res., 75, 1585-1613, 1970. 

Klerkx, J., S. Deutsch, and P. De Paepe, 
Rubidium, strontium content and 
strontium isotopic composition of 
strongly alkalic basaltic rocks from 
the Cape Verde Islands, Contrib. Min- 
eral. Petrol, 45, 107-118, 1974. 

Kudo, A. M., D. G. Brookins, and A. W. 
Laughlin, Sr isotopic disequilibrium in 
lherzolites from the Peurco Necks, 
New Mexico, Earth Planet. Sci. Lett., 
15, 291-295, 1972. 

Laughlin, A. W., and D. G. Brookins, 
Chemical and strontium isotopic in- 
vestigations of ultramafic inclusions 
and basalt, Bandera Crater, New 
Mexico, Geochim. Cosmochim. Acta, 
35, 107-113, 1971. 

Leggo, P. J., and R. Hutchison, A Rb-Sr 
isotope study of ultrabasic xenoliths 
and their basaltic host rocks from the 
Massif Central, France, Earth Planet. 
Sci. Lett, 5, 71-75, 1968. 

Lovering, J. F., and M. Tatsumoto, Lead 
isotopes and the origin of granulite 
and eclogite inclusions in deep-seated 
pipes, Earth Planet. Sci. Lett., 4, 350- 
356, 1968. 

Lussia-Berdou-Polve, M., and P. Vidal, 
Initial strontium isotope composition 
of volcanic rocks from Jan Mayen and 
Spitzbergen, Earth Planet. Sci. Lett., 
18, 333-338, 1973. 

Manton, W. I., and M. Tatsumoto, Some 
Pb and Sr isotopic measurements on 
eclogites from the Roberts Victor 
Mine, South Africa, Earth Planet. Sci. 
Lett., 10, 217-226, 1971. 

Meijer, A., Pb and Sr isotopic data bear- 
ing on the origin of lavas from the 
Mariana arc system, submitted to 
Bull. Geol. Soc. Am., 1975. 

Misener, D. J., Cationic diffusion in oli- 
vine to 1400°C and 35 K bar, in Geo- 
chemical Transport and Kinetics, A. 
W. Hofmann, B. J. Giletti, H. S. 
Yoder, Jr., and R. A. Yund, eds., 
Carnegie Inst. Wash. Publ. 634, PP- 
117-125, 1974. 

Modreski, P. J., and A. L. Boettcher, 
Phase relations of phlogopite in the sys- 
tem K 2 0-MgO-CaO-Al 2 3 -Si0 2 -H 2 
to 35 kilobars: a better model for micas 
in the interior of the earth, Am. J. Sci., 
273, 385, 1973. 

Morioki, M., and K. Kigoshi, Lead iso- 
topes and age of Hawaiian lherzolite 



nodules, Earth Planet. Sci. Lett., 25, 
116-120, 1975. 

Muehlenbachs, K., and I. Kushiro, Oxy- 
gen isotope exchange and equilibrium 
of silicates with C0 2 or 2 , in Carnegie 
Inst. Wash. Year Book 73, pp. 232- 
236, 1974. 

Mysen, B. 0., and A. L. Boettcher, Melt- 
ing of a hydrous mantle: II. Geo- 
chemistry of crystals and liquids 
formed by anatexis of mantle perido- 
tite at high pressures and high tem- 
peratures, as a function of controlled 
activities of water, hydrogen, and car- 
bon dioxide, J. Petrol., in press, 1975. 

O'Hara, M. J., Non-primary magmas 
and dubious mantle plume beneath 
Iceland, Nature, 243, 507-508, 1973. 

O'Hara, M. J., Is there an Icelandic man- 
tle plume? Nature, 253, 708-710, 1975. 

O'Neil, J. R., C. E. Hedge, and E. D. 
Jackson, Isotopic investigations of 
xenoliths and host basalts from the 
Honolulu volcanic series, Earth Planet. 
Sci. Lett., 8, 253-257, 1970. 

O'Nions, R. K., and K. Gronvold, Petro- 
genetic relationships of acid and basic 
rocks in Iceland: Sr isotopes and rare- 
earth elements in late and postglacial 
volcanics, Earth Planet. Sci. Lett., 19, 
397-409, 1973. 

O'Nions, R. K., and R. J. Pankhurst, 
Secular variation in the Sr-isotope 
composition of Icelandic volcanic 
rocks, Earth Planet. Sci. Lett., 21, 13- 
21, 1973. 

O'Nions, R. K., and R. J. Pankhurst, 
Petrogenetic significance of isotope 
and trace element variations in vol- 
canic rocks from the mid-Atlantic, J. 
Petrol, 15, 603-634, 1974. 

O'Nions, R. K., R. J. Pankhurst, I. B. 
Fridleifsson, and S. P. Jakobsson, 
Strontium isotopes and rare-earth ele- 
ments in basalts from Heimaey and 
Surtsey volcanic eruption, Nature, 
243, 213-214, 1973. 

Paul, D. K., Strontium isotope studies 
on ultramafic inclusions from Dreiser 
Weiher, Eifel, Germany, Contrib. 
Mineral. Petrol, 34, 22-28, 1971. 

Peterman, Z. E., I. S. E. Carmichael, and 
A. L. Smith, Strontium isotopes in 
quaternary basalts of southwestern 
California, Earth Planet. Sci. Lett., 7, 
381-384, 1970. 

Petrovic, R., The effect of coherency 
stress on the mechanism of the reac- 
tion albite -(- K+ ^± K-feldspar + 
Na+ and on the mechanical state of 
the resulting feldspar, Contrib. Min- 
eral. Petrol, 41, 151-170, 1973. 

Philpotts, J. A., and C. C. Schnetzler, 
Phenocryst-matrix partition coeffi- 
cients for K, Rb, Sr, and Ba, with 
applications to anorthosite and basalt 
genesis, Geochim. Cosmochim. Acta, 
34, 307-322, 1970. 

Richter, F. M., and B. Parsons, On the 
interaction of two scales of convection 
in the mantle, J. Geophys. Res., 80, 
2529-2541, 1975. 

Ringwood, A. E., The petrological evolu- 
tion of island arc systems, J. Geol 
Soc. London, 130, 183-204, 1974. 

Schilling, J.-G., Iceland mantle plume: 
geochemical evidence along Reykjanes 
Ridge, Nature, 242, 565-571, 1973. 

Seitz, M. G., Uranium and thorium dif- 
fusion in diopside and fluorapatite, in 
Carnegie Inst. Wash. Year Book 72, 
pp. 586-588, 1973. 

Shimizu, N., An experimental study of 
the partitioning of K, Rb, Cs, Sr and 
Ba between clinopyroxene and liquid 
at high pressures, Geochim. Cosmo- 
chim. Acta, 38, 1789-1798, 1974. 

Sigvaldason, G. E., Basalts from the 
center of the assumed Icelandic mantle 
plume, J. Petrol, 15, 497-524, 1974. 

Sinha, A. K., and S. R. Hart, A geo- 
chemical test of the subduction hy- 
pothesis for generation of island arc 
magmas, in Carnegie Inst. Wash. Year 
Book 71, pp. 309-312, 1972. 

Stuckless, J. S., and R. L. Ericksen, 
Strontium isotope geochemistry of the 
volcanic rocks and associated mafic 
and ultramafic nodules from the Ross 
Island area, Antarctica (Abs.) Eos, 

Stueber, A. M., and M. Ikramuddin, 



Rubidium, strontium in ultramaflc 
nodule minerals and host basalts, Geo- 
chim. Cosmochim. Acta, 38, 207-216, 

Stueber, A. M., and V. R. Murthy, 
Strontium isotope and alkali element 
abundances in ultramaflc rocks, Geo- 
chim. Cosmochim. Acta, 30, 1243-1259, 

Stull, R. J., and K. McMillan, Origin of 
lherzolite inclusions in the Malapai 
Hill basalt, Joshua Tree National 
Monument, California, Geol. Soc. Am. 
Bull, 84, 2343-2350, 1973. 

Subbarao, K. V., The strontium isotopic 
composition of basalts from the East 
Pacific and Chile Rises and abyssal hills 
in the eastern Pacific Ocean, Contrib. 
Mineral. Petrol, 37, 111-120, 1972. 

Subbarao, K. V., and C. E. Hedge, K, 
Rb, Sr and 87 Sr/ 86 Sr in rocks from the 
mid-Indian oceanic ridge, Earth Planet. 
Sci. Lett., 18, 223-228, 1973. 

Sun, S. S., and G. N. Hanson, Evolution 
of the mantle: geochemical evidence 
from alkali basalt, Geology, 3, 297- 
302, 1975. 

Sun, S. S., and B. Jahn, Lead and stron- 
tium isotopes in post-glacial basalts 
from Iceland, Nature, 255, 527-530, 

Tatsumoto, M., Genetic relations of oce- 
anic basalts as indicated by lead iso- 
topes, Science, 153, 1094-1101, 1966a. 

Tatsumoto, M., Isotopic composition of 
lead in volcanic rocks from Hawaii, 
Iwo Jima; and Japan, J. Geophys. Res., 
71, 1721-1733, 19666. 

Wilshire, H. G., and E. D. Jackson, 
Problems in determining mantle geo- 
therms from pyroxene compositions in 
ultramaflc rocks, J. Geol, 83, 313-329, 

Yoder, H. S., Jr., and I. Kushiro, Melt- 
ing of a hydrous phase: phlogopite, 
Am. J. Sci, 267 A, 558, 1969. 

Zartman, R. E., and F. Tera, Lead con- 
centration and isotopic composition in 
five peridotite inclusions of probable 
mantle origin, Earth Planet. Sci. Lett., 
20, 54-66, 1973. 

Kinetic Processes and Thermal 
History of Rocks 

Martin H. Dodson 


Kinetic processes in minerals and 
rocks have been little used for geo- 
thermometry, in spite of their very great 
sensitivity to temperature. The scarcity 
of reliable kinetic data applicable to 
geologic situations has been partly re- 
sponsible for this state of affairs. An- 
other factor is the lack of an adequate 
body of theory to relate kinetic processes 
to thermal history. In view of the rapid 
increase in quantity and quality of 
kinetic data now becoming available, it 
appears timely to review some theoreti- 
cal relationships related to kinetics in 
slowly cooling systems, and to present 
some new relationships concerning ki- 
netic processes during thermal pulses 
produced by igneous intrusions in their 
host rocks. The application of these 
relationships both to loss of radiogenic 
daughter products and to false petro- 
chemical equilibria will be discussed. 

Published Work 

Gentner, Goebel, and Prag (1954) 
solved the differential equations for 
simultaneous production and diffusion of 
radiogenic daughter product in a cooling 
cubic crystal ; they used the result in in- 
ferring a cooling history for the Stass- 
furt salt deposits from K-Ar and U-He 
ages. Wagner and Reimer (1972) com- 
bined data on fission-track annealing 
with fission-track ages to infer the low- 
temperature cooling history of the Alps. 
Dodson (1973) discussed the concept of 
closure temperature in cooling geochron- 
ological and petrological systems, and 
found a simple relationship between it 
and cooling rates. Wood (1964) and 
Goldstein and Short (1967) calculated 
cooling rates of iron meteorites by nu- 
merical modeling of the nickel concen- 
tration profiles across taenite lamellae. 
Lasaga and Richardson (1975) found 



analytic solutions for diffusion-controlled 
exchange between adjacent mineral 
phases in a cooling metamorphic assem- 
blage, and applied them to Fe-Mg ex- 
change between garnet and cordierite; 
their method permits the use of various 
temperature-time relationships during 
cooling. Seifert and Virgo (1975) esti- 
mated the cooling rate of a natural 
anthophyllite from the Fe-Mg order- 
disorder relationships. Mineral ages in 
contact zones have been studied by Hart 
(1964) and Hanson and Gast (1967), 
among others, using numerical methods 
to model the kinetics of isotopic losses. 
Brandt (1974) made the surprising 
claim, backed by theoretical analysis, 
that it is possible to infer temperatures 
in contact zones from isotopic losses 
without a knowledge of the parameters 
governing the rate of loss ; a partial anal- 
ysis of this claim is presented below. 

Approximations to Thermal History 

At sufficiently high temperatures any 
given kinetic process will go so rapidly, 
in relation to the rate of change of tem- 
perature, that equilibrium will be con- 
tinuously maintained. At sufficiently low 
temperatures, the process will go so 
slowly that no change can occur over 
geological time. In neither of these situ- 
ations will kinetics tell us anything 
about thermal history except, perhaps, 
lower or upper limits to temperatures. 
However, in the transitional temperature 
range between those two extremes, ki- 
netic processes will go part way towards 
equilibrium, and the system under con- 
sideration will then contain a quantita- 
tive record of some aspect of that part of 
the thermal history. 

The transitional temperature range is 
fairly narrow because of the tempera- 
ture sensitivity of kinetic processes. Over 
this limited range it generally will be 
appropriate to describe the rate/tem- 
perature relationship by the Arrhenius 

k(T) = fc exp (-E/RT) (1) 

where k(T) is the rate constant at ab- 
solute temperature T; R is the gas con- 
stant; k is the frequency factor; and E, 
the activation energy characteristic of 
the process. For volume diffusion we re- 
place k{T) and fc„ by D(T)/a 2 and 
D () /a 2 , a being a characteristic dimen- 
sion of the system. 

The form of the Arrhenius relation- 
ship makes it desirable to describe ther- 
mal history in terms of the reciprocal of 
absolute temperature, which can then be 
expressed as a Taylor series in time 
t, i.e., 

T-Ht)=T ~i + 


(t-t ) 


d^T- 1 (t-t ) 2 




in which T'o and the differential coeffi- 
cients are taken at time t = ft,. Because 
we are concerned with the limited tran- 
sitional temperature range, we can con- 
fine our attention to the following ap- 

(1) Slow cooling: 

dT- 1 


= T -i_ T 

o • (t-T )/T 2 

(2) Thermal pulse: 
T~ l ~ T- 1 


d 2 T (t- t ) 2 



dt 2 2! 


T(t — t )-/2T 2 


Equation 4 corresponds to a parabolic 
approximation symmetrical about the 
peak of the pulse at t = t . 

In all cases investigated it is necessary 
that the kinetic processes under consid- 
eration be so slow at ambient tempera- 
tures as to cause negligible changes over 
geological periods. 



Slow Cooling 

The closure of a cooling penological 
system is illustrated in Fig. 49. The or- 
dinate x represents a chemical parame- 
ter (e.g., 18 abundance, or an Fe 2 +-Mg 
ordering parameter) whose equilibrium 
value varies directly with temperature 
according to curve A. In the high tem- 
perature range equilibrium is maintained 
continuously, and the actual value of x, 
represented by curve B, follows curve A 
very closely. In the transitional range 
curve B departs from curve A, though x 
still changes toward the equilibrium 
value. Eventually, curve B asymptoti- 
cally approaches a constant value of x, 
which corresponds to the observed clo- 
sure temperature T c . 

For radiogenic isotopes an analogous 
situation exists (Dodson, 1973). Curve 
A corresponds to the total amount of 
radiogenic daughter produced up to any 
time, and curve B to the deficit of radio- 
genic daughter product in the mineral, 

Time — >• 




Low temp. 

High i 
temp. | 
range | 



/ i 



B s/ 



Fig. 49. Definition of closure temperature T c . 
A, equilibrium value of X. B, actual value of 
X. C, Accumulation of radiogenic isotope 

which is the difference between curve A 
and curve C. The important difference 
between geochronological and petro- 
chemical systems is that in the former 
we must estimate T c from kinetic pa- 
rameters and cooling rates, whereas in 
the latter we can directly measure T c . 
Differential equations for curve B 
have been solved by Dodson (1973) for 
two simple cases, namely, first-order 
chemical reaction, and volume diffusion 
of a minor component between a geo- 
metrically simple isotropic body and an 
infinite, mobile reservoir. In both cases 
the boundary conditions (represented by 
curve A) are assumed to vary linearly 
with time, and therefore with 1/T, over 
the transitional temperature range. The 
cooling rate (T 7 ) is taken into account in 
a cooling time constant r, which is de- 
fined by a dimensionless form of Equa- 
tion 3 




RT + r 

By differentiating this we find 
r = -RT 2 /ET 
From Equations 1 and 4 we find 
k(t) = /c(0)e~</ T 




so t is the time required for the rate con- 
stant to diminish to 1/e of any given 

Use of Equation 7 makes tractable 
the integration of the appropriate linear 
differential equations, leading to the fol- 
lowing results: 


RT ( 

In (Arko) 



= In (ArDo/a 2 ) 



in which A is 1.8 for first-order kinetics, 
and 8.7, 27, or 55 for volume diffusion 
from a plane slab, cylinder, or sphere, 

Combining Equations 6 and 8 we ob- 



= In [—ART* (k , D /a 2 )/ET] (9) 

which can be solved iteratively for T c or 
E, and directly for T or frequency fac- 
tor. In the former case a rather precise 
result is obtained from relatively impre- 
cise values of cooling rate; conversely, 
rather large errors in T, k , etc. would 
result from relatively small errors in T c 

For geochronological systems in young 
orogenic belts a precise thermal history 
can in principle be obtained. For a given 
locality, each system provides an equa- 
tion relating T c and T at the time cor- 
responding to the measured age. In the 
simplest case, a single system together 
with the present-day temperature will 
yield a linear approximation to cooling 
history. For n systems an nth degree 
polynomial in time could be fitted to the 
data, in principle. 

Use of geochemically observed values 
of T c (e.g., apparent 18 paleotempera- 
tures) to determine cooling rate will be 
less satisfactory. Very precise observa- 
tions of T c are needed; the activation 
energy must also be known very well; 
and errors in A or frequency factor will 
be propagated directly to the estimated 
cooling rate. It may be more satisfac- 
tory to use Equation 9 to compare cool- 
ing rates derived from a given kinetic 
system, rather than to determine abso- 
lute values. 

More Complex Kinetics 

Examination of the differential equa- 
tions for closure of cooling systems sug- 
gests that solutions will only take the 
form of Equation 8 with constant A if 
the differential equation is linear. How- 
ever, Dodson (19756) shows that Equa- 
tion 8, with a suitably chosen value of 
A, may well be a useful approximation 
for more complex kinetic systems — pro- 
vided the rate constant has dimensions 
1/time, the value of A is given within a 
factor of about 2 by 2/kt 1/2 , where t 1/z 
is the time required for 50% approach to 

equilibrium at constant temperature 
with rate constant k. For the order- 
disorder kinetics studied by Seifert and 
Virgo (1974, 1975) this approach yields 
A = 10. Given their observed T c of 
270°C, E = 61.6 kcal/mole, and k = 
8.6 X 10 17 yr _1 , we obtain a cooling rate 
of 13°C per million years. Numerical 
solution of the rate equations is in prog- 
ress and should indicate how good this 
simple approach is and to what extent A 
is a function of r k . It should be noted 
that in second or higher-order kinetics 
the quantity kt 1/2 is not uniquely de- 
fined as it is in first-order kinetics. 

Kinetic Effects of a Thermal Pulse 

In a zone surrounding any igneous 
intrusion the rocks are subjected to 
transient heating. The peak temperature 
reached during the heating falls off 
away from the intrusion, and the time 
of the heating tends to increase. A sim- 
ple approximation to the effective heat- 
ing time of such a thermal pulse will be 
derived here. 

In studying diffusion phenomena with 
time-varying diffusion coefficients, stand- 
ard solutions of the diffusion equation 
can be used simply by replacing the 
dimensionless time Dt/a 2 by 

(I /a 2 ) f*D{u)du. 

'a 2 )jl 

In general, for any kinetic effect expres- 
sible in the form 

x — k(t)f (x, t) 


in which k(t) has dimensions 1/time, a 
solution can be obtained in which the 
dimensionless parameter f* k (u) du takes 
the place of time. J 

For a thermal pulse we can define its 

effective duration t p as follows 



JD (t)dt 


where D max is the value of the diffusion 
coefficient at the maximum temperature. 
The limits of integration are not well 



defined, but they are not critical pro- 
vided the process is slow enough to be 
geologically negligible at ambient tem- 

Assuming a rapid diminution in the 
rate coefficient with diminishing tem- 
perature, we can express D(t) as a func- 
tion of time by using the Taylor ap- 
proximation given in Equation 4. We 
thus have 

so that : 



~ D max /exp 
E tfT^it—to) 2 


= D max ^-2RTlJET 

I e~ u ~ du 

Integrating from — oo to oo, we ob- 

f D(t)dt 

r p ~ V -2tRTIJET (13) 

V-2*RTiJET (12) 

The approximation upon which Equa- 
tion 13 is based is illustrated in Fig. 50. 
It has been tested by numerical integra- 
tion over the thermal history shown 
(which is the pulse shape due to an 
instantaneous plane source of heat at 
x = 0, t = 0) . Selected results are shown 
in Table 7. The magnitude of the error 
in the approximation can be seen to 
range from ~100% at E — 10 kcal/ 
mole to ~10% at E = 60 kcal/mole. 
Changing pulse height has little effect 
on these errors. (Very small errors were 
found in previous numerical tests quoted 
by Dodson (1975a) and are probably 
due to modeling a pulse symmetrical 
about T max , in which case the third- 
order term in the Taylor expansion is 
zero.) For many purposes this degree of 
approximation will be sufficient; how- 
ever, an improved approximation is 
being sought. 

Finite difference approximations to 
d 2 T/dt 2 can be found. The simplest 
second-difference approximation is T = 



Fig. 50. A, thermal pulse due to an instantaneous plane source. B, Taylor approximation to 
A. C, equivalent square pulse for E — 30 kcal/mole (See Table 7). 



TABLE 7. Relation between Effective Pulse Duration r v Calculated by 

Equation 20 and by Numerical Integration, for Pulse Shape 

AT = Br 1/2 e~ xlt with T a = 300, B - 600 

Activation Energy 

T P 

(Equation 20) 

r p 

(numerical integration) 




* The numerical integral converges very slowly for E =. 10 kcal/mole. 
Also the rate constant at ambient temperature is too large relative to that 
at T m ax (290°C) to be neglected over all but very short geological periods. 

4a7Va£ 2 where At is the duration meas- 
ured at half the maximum pulse height 
AT (Fig. 50) . This gives too low a value 
for the strongly asymmetrical pulse of 
Fig. 50, for which we find empirically 
T ~ 9AT/At 2 . In general we can write 

T = S P AT/At 2 , 


S p being a constant dependent upon 
pulse shape. 

We can now derive a simple expres- 
sion for the peak temperature which 
must be reached to attain a given ap- 
proach towards equilibrium, say 50%- 
We can call this the threshold tempera- 
ture T t h and write 

^"50 = "max Tp/a- 

= D exp (- E/RT th )r p /a 2 (15) 

where GV )0 is a factor dependent on the 
diffusion geometry, and is equal to 0.2, 
0.06, and 0.03 for planar, cylindrical, 
and spherical geometry (see Crank, 
1956) . We can thus write 




= IMG- 1 r p DJa-) 


mal models or indirectly from the dis- 
tribution of apparent ages. 

Diffusion of components other than 
radiogenic isotopes may be treated in a 
similar way. Although the equilibrium 
conditions are in general temperature 
dependent, and therefore will change 
with time, they can reasonably be re- 
garded as constant to a first approxima- 
tion during the significant period near 
the top of the pulse. Initial conditions 
would have to be determined outside the 
contact zone. For diffusion processes 
some knowledge of the initial concentra- 
tion profile of the diffusing component 
will be needed. 

Brandt's Method for Thermal Pulses 

Brandt (1974) suggested that tem- 
peratures in contact zones could be in- 
ferred from diffusive losses of radiogenic 
isotopes without prior knowledge of the 
diffusion parameters. Here the results 
so far obtained in a critical study of 
Brandt's method (which is not yet com- 
plete) will be summarized. 

Brandt assumes the thermal pulse to 
be of constant duration r, the peak 
height AT{x) being given by 

The similarity to Equation 8 for slow 
cooling is noteworthy. One may sub- 
stitute for r p by Equations 13 and 14, 
and solve iteratively for 7\h, the result 
being fairly insensitive to the argument 
of the logarithmic expression. 

It may well be possible to make ade- 
quate estimates of At directly from ther- 

aT(x) = AToe- 1 " 


where x is the distance from the contact; 
k and AT are undetermined. He then 
substitutes Equation 17 in the Arrhenius 
equation, to express the quantity Dr/a 2 
(which he calls the Fourier number Fo) 
as a function of x; this quantity is in- 



ferred directly from the observed diffu- 
sion losses, assuming a simple model. 
Brandt then shows that the natural 
logarithm of Fo plotted against x will 
have a point of inflection at distance x inf 
at which the absolute temperature T in f 
is twice the ambient temperature T a . Two 
further observed quantities, namely, the 
value of d{\nF )/dx at the point of in- 
flection, and the ratio of F at that point 
to F at the contact, make it possible to 
determine the parameters AT and k, de- 
fining the pulse height as a function of x, 
and the activation energy E. 

It is not easy to give a physical inter- 
pretation of Brandt's rather elaborate 
mathematical procedure. However, one 
may make the general comment that it 
is an attempt to combine the laws gov- 
erning diffusion at particular locations 
with those governing the transfer of heat 
among those locations. 

An attempt has been made to gen- 
eralize Brandt's method to other tem- 
perature profiles. If we write 

= -An 


AT = ATof(kx) 


where f(kx) is a known function of x 
except for the single unknown parameter 
k, we can proceed along parallel lines to 
Brandt and find the point of inflection 
in the curve of Fourier number versus 
distance. This will yield the following 



= (l 


= (2/'vr - /)._._ d9b) 


Equation 19b relates the two unknowns 
aT and k to the known T a and x int . Two 
additional equations needed analogous 
to Brandt's Equations 11 and 12 are: 



Foi E 

Fo 2 ~~ RT~ 2 — RJ\ 

_1_ dFo d(lnFo) 
Fo dx dx 


RT 2 t 


kf'{kx) (21) 

in which T inf , T 1} and T 2 are related by 
Equation 18 to AT , k, and the known 
distances x int , x lf and x 2 . The distances 
x x and x 2 would be selected to yield the 
most precise observable estimate of 
ln(Foi/Fo 2 ). (The quantity Fo at the 
contact, used by Brandt, will generally 
be indeterminate, since total loss of 
radiogenic isotopes normally occurs 
there at the time of intrusion.) 

Equations 19b, 20, and 21, in conjunc- 
tion with Equation 18, are sufficient in 
principle to solve the problem. The two 
questions to be answered are: "How sen- 
sitive is the result to errors in the obser- 
vations?" and "How sensitive is the 
method to the assumed form of /(/ex)?" 
To the first question one may answer, 
qualitatively, that the position of the 
point of inflection and the gradient of 
the curve at that point will not be deter- 
minable very precisely, either by graphi- 
cal or numerical methods; besides errors 
of measurement, there are inhomogenei- 
ties in the heated rock to be considered. 
The latter can be seen clearly in the 
results of Hanson and Gast (1967) on 
the Snowbank Lake granite in the au- 
reole of the Duluth gabbro. To answer 
the second question, one can try one or 
two simple tests. For example, if we use 
instead of Equation 18 

AT oc (kx) n 

in which n is negative, we find on differ- 
entiating and substituting in Equation 
19a that T a /T int = (1 + l/n)/2. Now 
for the temperature profile due to an 
instantaneous plane source (see Table 
7) we find n = — 1, so no point of inflec- 
tion could be observed. Taking n = — 2, 
we get T inf = 47V These results suggest 
that the calculation will be very sensi- 
tive to the form oij(kx) . 

Brandt's assumption that the pulse is 
of constant duration is rather unsatis- 
factory. Using the approximation of 
Equation 13 and the pulse shape given in 



Table 7, we can show that the effective 
pulse duration varies approximately as 
x 2 . However, if we knew the precise 
form of the temperature profile we would 
presumably know how the effective pulse 
duration varies, so we could, in princi- 
ple, derive rather more elaborate ver- 
sions of Equations 19b-21. 

As an alternative to the procedure dis- 
cussed above we may suppose that we 
have a thermal model in which the linear 
dimensions of the contact zone are known 
in terms of the thermal properties of the 
rock and the size of the intrusion caus- 
ing the thermal pulse. Then we can write 
AT = AT f(P, kx), in which k is known, 
and P is some other parameter of the 
thermal model (e.g., latent heat, or a 
parameter describing the effect of fluids 
on heat transfer in the contact zone). 
Then P and AT Q would be determinable, 
in principle. 

Conclusions and Warning 

The various approaches to cooling 
history and to thermal pulses discussed 
here are idealizations. While they pro- 
vide a guide to the study of the rela- 
tionship between kinetic processes and 
thermal history, the uncertainties of 
extrapolation of kinetic data to low tem- 
peratures, particularly in the mecha- 
nisms operative over long periods at low 
temperatures, call for great caution in 
interpretation. It is reasonable to hope, 
however, that by testing these relation- 
ships on a number of different kinetic 
systems in various environments we shall 
discover eventually those systems which 
are most promising for use in geo- 


Brandt, S. B., A new approach to the 
determination of temperatures of in- 
trusives from radiogenic argon loss in 
contact aureoles, in Geochemical 
Transport and Kinetics, A. W. Hof- 
mann, B. J. Giletti, H. S. Yoder, Jr., 
and R. A. Yund, eds., Carnegie Inst. 
Wash. Publ. 634, PP- 295-298, 1974. 

Crank, J., Mathematics of diffusion, Ox- 
ford, Clarendon Press, 1956. 

Dodson, M. H., Closure temperature in 
cooling geochronological and peno- 
logical systems, Contrib. Mineral. 
Petrol., 40, 259-276, 1973. 

Dodson, M. H., Diffusion effects of a 
thermal pulse (Abstr.), Eos, 56, 472, 

Dodson, M. H., Arrhenius processes in 
cooling solids, in preparation, 19756. 

Gentner, W., K. Goebel, and R. Prag, 
Vergleichende Messungen nach der 
Kalium-Argon und Uran-Helium Me- 
thode, Geochim. Cosmochim. Acta, 5, 
125-136, 1954. 

Goldstein, J. I., and J. M. Short, Cooling 
rates of 27 iron and stony-iron mete- 
orites, Geochim. Cosmochim. Acta, 31, 
1001-1023, 1967. 

Hanson, G. N., and P. W. Gast, Kinetic 
studies in contact metamorphic zones, 
Geochim. Cosmochim. Acta, 31, 1119- 
1153, 1967. 

Hart, S. R., The petrology and isotopic 
mineral age relations of a contact zone 
in the Front Range, Colorado, J. Geol., 
72, 493-525, 1964. 

Lasaga, A. C, and S. M. Richardson, A 
mathematical model of diffusion in 
retrograde metamorphism (Abstr.), 
Eos, 56, 459, 1975. 

Seifert, F., and D. Virgo, Temperature 
dependence of intra-crystalline Fe 2+ - 
Mg distribution in a natural antho- 
phyllite, Carnegie Inst. Wash. Year 
Book 73, pp. 405-411, 1974. 

Seifert, F., and D. Virgo, Kinetics of the 
Fe 2 + -Mg order-disorder reaction in 
anthophyllites ; quantitative cooling 
rates, Science, 188, 1107-1109, 1975. 

Wagner, G. A., and G. M. Reimer, Fis- 
sion track tectonics: the tectonic inter- 
pretation of fission track apatite ages, 
Earth Planet. Sci. Lett., 14, 263-268, 

Wood, J. A., The cooling rate and parent 
planets of several iron meteorites, 
Icarus, 3, 429-459, 1964. 



Deep Sea Drilling Project Leg 34 

Basalts and Ocean Ridge Basalts: 

A Geochemical Comparison 



Basalt produced at mid-ocean ridges 
(MORB) has been recognized for sev- 
eral years as being distinctly different 
from basalts produced in other tectonic 
environments. While the differences are 
small, they are nevertheless adequate for 
distinguishing MORB from other basalts, 
and there is currently considerable de- 
bate over the explanation of these differ- 
ences. While the geographic variation of 
MORB chemistry is well studied, largely 
from dredge hauls along the ridges, the 
variation back in time is only poorly 
understood. Older MORB usually are 
accessible only by drilling, and fre- 
quently this older material is badly al- 
tered. Leg 34 of the Deep Sea Drilling 
Project (DSDP) drilled three sites on 
the Nazca plate, between the East Paci- 
fic Rise and the Peru-Chile trench. The 
basalts drilled on this leg, covering an 
age range of about 15-40 m.y., are less 
altered than most DSDP basalts and 
provide an ideal suite of samples to 
compare with zero-age MORB. This 
paper will present such a comparison, 
utilizing major elements, the LIL (large- 
ion lithophile) trace elements, and Sr 
isotope ratios. 

Technique and Results 

Major element analyses were per- 
formed by microprobe on glass samples 
from sites 319 and 320, leg 34; no glass 
was available from site 321. 

Analyses of K, Rb, Cs, Sr, Ba, and the 
rare earth elements (REE) were carried 
out on a few selected samples by isotope 
dilution, and Sr concentration on a com- 
plete suite of leg 34 samples was deter- 
mined by x-ray fluorescence. The Sr iso- 
topic ratios were performed on a 6-inch 
automated mass spectrometer. In addi- 
tion to the leg 34 samples, LIL-element 

and Sr isotopic data on a large number 
of young MORB dredge samples and 
various oceanic island and island arc 
samples are reported for comparison 


Sr concentrations. Sr is relatively in- 
sensitive to alteration by seawater (Hart, 
1969; Pearce and Cann, 1973; Thomp- 
son, 1973) and so may be used for the 
leg 34 basalts to characterize inter- and 
intra-site variations. The data are plotted 
as a function of depth at the three sites 
in Fig. 51. While there is some overlap 
between sites, each site has a relatively 
distinct range of Sr concentration, with 
site 320 having the highest average con- 
centration (126 ppm) ; site 321, the low- 
est (92 ppm) ; and site 319, an inter- 
mediate value (111 ppm). At site 320 
there is a slight tendency for Sr concen- 
tration to increase with depth; the other 
two sites do not show any clear trend 
with depth. Even single cooling units 


Sr concentration, ppm 
IOO 110 120 130 



60 L- 

Fig. 51. Strontium concentration versus 
depth, sites 319, 320, and 321, leg 34. Precision 
estimate of ±3% shown at bottom. 



may show considerable variation, as in 
the unit at site 319 between 11 m and 
30 m. While most of this unit shows 
values in the range 105-115 ppm, three 
samples fall above 125 ppm. These high 
values may be related to local accumula- 
tion of plagioclase phenocrysts. Overall, 
the Sr concentrations of leg 34 basalts 
are within the range of typical MORB, 
and significantly lower than Sr concen- 
trations found in subalkaline tholeiites 
from oceanic islands or island arcs where 
concentrations are typically greater than 
200 ppm. In relatively fresh rocks, Sr 
concentrations can be used to discrimi- 
nate MORB from basalts from almost 
all other environments (one exception 
being Iceland, where low-Sr tholeiites 
are also found, Hart et al., 1973). 

Rare earth element patterns. Of the 
various LIL elements, the REE are the 
most insensitive to seawater alteration 
(Philpotts et al, 1969). The REE pat- 
terns for glass samples from sites 319 
and 320 are shown in Fig. 52, along with 
two samples of MORB which represent 
the upper and lower limits of all MORB 
analyzed to date (Kay et al., 1970). 
The site 319 and 320 patterns show the 
light REE depletion typical of most 
MORB, with no significant europium 

anomaly. There is a tendency for deple- 
tion of the heavy REE in these two leg 
34 basalts, and this is not common in 
MORB. REE patterns of the type shown 
in Fig. 52 are not diagnostic of MORB, 
since light-depleted patterns may also 
be found in tholeiites from island arcs 
and some oceanic islands such as Iceland. 
K, Rb, Cs, and Ba concentrations. 
These elements are all susceptible to sea 
water alteration, with Cs being one of 
the most strongly affected of all ele- 
ments. However, in fresh glasses such as 
those from sites 319 and 320, the alkali 
data should be reliable (Table 8). The 
sample from site 321 was not a glass, 
but a holocrystalline sample with rela- 
tively low water content. The low Cs 
content of this sample suggests it is 
relatively free from alteration effects. 
The samples from all three sites are low 
in alkalies and Ba, and show the high 
ratios of K/Rb, K/Cs, and K/Ba that 
are typical of MORB. In fact, the site 
319 glass is lower in K (382 ppm) than 
any other MORB thus far analyzed by 
accurate techniques. There appears to 
be a trend of increasing K and Ba con- 
centration (and decreasing K/Rb, K/Ba 
ratio) with age represented by the three 
samples of Table 8 (site 319, 20 m.y.; 
















Fig. 52. Rare earth element patterns for two glass samples, 319A-2-1, 27-30, and 320B-3-1, 
101-105. Shown for comparison are the highest and lowest patterns (dashed lines) reported 
for MORB, from Kay et al, 1970. Note the lack of significant europium anomaly in the leg 
34 samples. 



TABLE 8. LIL Element Contents, Leg 34 Basalts 












K (ppm) 






Rb (ppm) 






Cs (ppm) 






Sr (ppm) 

95 (XRF) 



86 (XRF) 


Ba (ppm) 
























87 Sr/ 8e Sr + 






H 2 (total, %) 




H 2 Of (%) 


* Average of 22 fresh basalts from four different ridges (Hart, 1975). 
t Precision 2<r = ±.01% relative to 0.70800 E + A standard. 

site 320, 30 m.y. ; site 321, 40 m.y.) ; more 
analyses would be required to see if 
this effect is real. 

To contrast the K, Rb, and Ba data 
of leg 34 basalts with other MORB and 
basalts from other tectonic environ- 
ments, concentration of these elements 
has been plotted using the data of Table 
8 and data from Hart (1975) in Figs. 53 
and 54. In the K/Rb-K plot (Fig. 53), 
site 319 falls slightly outside the field of 
MORB (due to the low K content); 
sites 320 and 321 fall in the central part 
of the MORB field. In general, MORB 
cannot be distinguished from arc tholei- 
ites on the basis of K-Rb relationships. 
The leg 34 basalts do fall outside the 
arc tholeiite field, though they show the 
same general trend of decreasing K/Rb 
with increasing K content that is shown 
by arc tholeiites, MORB, and the Reyk- 
janes Ridge-Iceland suite. 

In the K-Ba plot (Fig. 54), site 319 
again falls outside the MORB field 
(showing the highest K/Ba ratio yet 
recorded for any basalt), while sites 320 
and 321 fall within the MORB field. The 
K/Ba ratio appears to be an excellent 
diagnostic parameter (in fresh samples) 
for distinguishing MORB from arc 
tholeiites and oceanic island tholeiites 

(such as Iceland and Hawaii). While the 
leg 34 samples show a trend of negative 
slope on this plot (as they did in Fig. 
53), there is not as clear a trend in the 
arc tholeiite data. The Reyk janes 
Ridge-Iceland data show a trend of 
negative slope similar to the leg 34 sam- 
ples. Whether this is a trend which re- 
flects mantle chemistry directly, or is a 
result of differentiation or mixing proc- 
esses, is unknown at this time. The 
Reykjanes Ridge-Iceland trend has been 
interpreted by Schilling (1973) as mix- 
ing of two chemically different types of 
mantles (or of magmas from two differ- 
ent mantles). 

Sr isotope ratios. The data for the 
three leg 34 samples (Table 8) are com- 
pared in Fig. 55 with data from other 
MORB (Hart, 1975) and with data on 
basalts from other tectonic environments 
(unpublished data). On this plot, all 
three leg 34 sites fall within the MORB 
field. As has been noted before (Tatsu- 
moto et al.j 1965; Hart, 1971 ; Hart et al, 
1973), MORB generally have 87 Sr/ 86 Sr 
ratios lower than 0.7030, whereas most 
other basalts have ratios higher than 
0.7030. There are several exceptions: 
Basalts from the mid-Atlantic ridge at 
45°N have ratios of 0.7033 (Hart, 1975), 








I I I 

a Iceland, alkali basalts 
D Hawaii , tholeiites 

Inter-arc basins 

Arc basalts 


200 500 1000 2000 5000 10,000 


Fig. 53. K/Rb ratio versus K concentration of three leg 34 basalts compared with zero-age 
MORB (mid-ocean ridge basalt); island arc tholeiites; tholeiites from the Reykjanes Ridge, 
Iceland, and Hawaii; and five tholeiites from the marginal sea behind the Marianna Island 
arc. The MORB data is from Hart (1975); samples from 46°N on the mid-Atlantic ridge 
have not been plotted. Most of the other data are unpublished. The Iceland sample with very 
low K content is a Mg-rich basalt from the Reykjanes Peninsula. 










i •\*r^ Inter- arc basins 

Arc basalts 

A A 

a Iceland, alkali basaltsl 
° Hawaii , tholeiites 

200 500 1000 2000 5000 10,000 


Fig. 54. K/Ba ratio versus K concentration of three leg 34 basalts compared with basalts 
from other tectonic environments. See Fig. 53 legend for additional details. 












Hawaiian ^J°a °^Arc basalts 




basins — 






Fig. 55. 87 Sr/ 86 Sr ratio versus K/Rb ratio for three leg 34 basalts compared with basalts from 
other tectonic environments. See Fig. 53 legend for additional details. 

and several basalts from the Indian 
Ocean have ratios in the range 0.7030- 
0.7034 (Hart, 1975; Hedge et al, 1973). 
In addition, several oceanic islands have 
values as low as 0.7028 (St. Helena, 
Grant et al, 1974; Ascension, O'Nions 
and Pankhurst, 1974). Oceanic island 
basalts generally have K/Rb < 600, so 
in a plot like Fig. 55 the MORB fields 
and oceanic island fields remain sepa- 
rated (except for the peculiar u unde- 
pleted" basalts at 45°N on the mid- 
Atlantic ridge). It should be noted that 
there is a real variation in the 87 Sr/ 86 Sr 
of MORB from different localities, with 
well-documented values as low as 0.7023 
(Fig. 55) , contrary to the statement by 
O'Nions and Pankhurst (1974) that 
MORB tend to be very uniform at a 
value of 0.7029. An exceptionally low 
value of 0.7012 was reported by Hedge 
and Peterman (1970) for MORB from 
the Gordo Rise; reanalysis of powder 
from the same sample at DTM gave a 
more normal value of 0.7024. 

The differences between Sr isotope 
ratios in MORB and other basalts (for 
example, those from oceanic islands) 
have been ascribed to derivation from a 
mantle which is chemically heterogene- 
ous (Tatsumoto et al, 1965). Further 
work, in particular a detailed study of 
basalts from Iceland and the Reykjanes 
Ridge (Hart et al., 1973) has been simi- 
larly interpreted. The only serious alter- 
native to this heterogeneous mantle 
model is the one proposed by O'Nions 
and Pankhurst (1974) which invokes 
isotopic disequilibrium or heterogeneity 
on a mineralogical scale. Because of the 
isotopic differences between basalts from 
the three leg 34 sites, they would nor- 
mally be interpreted as being derived 
from chemically different mantle. In 
other words, many of the major and 
trace element differences between the 
leg 34 basalts may be primary, and a 
reflection of chemical differences in the 
mantle which supplied each of the sites. 
While some of the observed differences 



may be a result of crystal fractionation 
effects, there would be no need to explain 
all the differences in this way. Whether 
or not one can also explain the trace 
element and isotopic differences in terms 
of a mineral disequilibrium model (such 
as that of O'Nions and Pankhurst, 1974 ) 
must await further quantitative devel- 
opment of such models. Detailed argu- 
ments against disequilibrium melting 
models are given in a companion paper 
in this Report (Hofmann and Hart). 

Alteration effects. One test was made 
of the effect of seawater alteration on a 
drilled basalt. A core fragment from 
site 319A (see Table 8), which contained 
both a chilled glass margin and a holo- 
crystalline interior, was analyzed. 

The K content of the altered interior 
is 5 times and the Rb content 25 times 
that of the fresh glass (despite a water 
content for the interior sample of only 
0.44%), and the K/Rb and K/Cs ratios 
are drastically lowered (factors of 5 in 
both cases). The Ba concentration has 
doubled in the altered interior, while the 
Sr concentration is not significantly 
changed. The Sr isotope ratio is con- 
siderably higher in the altered portion 
(0.7031 versus 0.7027), presumably due 
to exchange with seawater Sr (Hart et 
al., 1974) which has a 87 Sr/ 86 Sr ratio 
of about 0.7091. In general, these effects 
are similar to those reported in other 
studies (Hart, 1969; Philpotts et al., 
1969; Thompson, 1973; Hart et al, 
1974) except that the interior portion 
of this sample would not be judged to 
be highly altered on the basis of its 
H 2 content alone. 


Grant, N. K., J. L. Powell, J. V. Walther, 
and F. R. Burkholder, The isotopic 
composition of strontium in lavas from 
St. Helena, South Atlantic, submitted 
to Earth Planet. Sci. Lett., 1974. 

Hart, S. R., K, Rb, Cs content and 
K/Rb, K/Cs ratios of fresh and al- 
tered submarine basalts: Earth Planet. 
Sci. Lett., 6, 295, 1969. 

Hart, S. R., K, Rb, Cs, Sr, and Ba con- 
tents and Sr isotopic ratios of ocean 
floor basalts: Philos. Trans. R. Soc. 
London, Ser. A, 268, 573, 1971. 

Hart, S. R., LIL-element geochemistry, 
leg 34 basalts: Initial Reports of the 
Deep Sea Drilling Project, 34, in press, 

Hart, S. R, J. G. Schilling, and J. L. 
Powell, Basalts from Iceland and along 
the Reykjanes Ridge: Sr isotope geo- 
chemistry, Nature, Phys. Sci., 246, 104, 

Hart, S. R, A. J. Erlank, and E. J. D. 
Kable, Sea floor basalt alteration: 
Some chemical and Sr isotopic effects, 
Contrib. Mineral. Petrol., 44, 219, 

Hedge, C. E., and Z. E. Peterman, The 
strontium isotopic composition of ba- 
salts from the Gordo and Juan de Fuca 
Rises, Northeastern Pacific Ocean, 
Contrib. Mineral. Petrol., 27, 114, 

Hedge, C. E., N. D. Watkins, R. A. 
Hildreth, and W. P. Doering, Sr 87 /Sr 86 
ratios in basalts from islands in the 
Indian Ocean, Earth Planet. Sci. Lett., 
21, 29, 1973. 

Kay, R., N. J. Hubbard, and P. W. 
Gast, Chemical characteristics and 
origin of oceanic ridge volcanic rocks, 
J. Geophys. Res., 75, 1585, 1970. 

O'Nions, R, K., and R. J. Pankhurst, 
Petrogenetic significance of isotope 
and trace element variations in vol- 
canic rocks from the Mid- Atlantic, 
J. Petrol. 15, 603, 1974. 

Pearce, J. A., and J. R. Cann, Tectonic 
setting of basic volcanic rocks deter- 
mined using trace element analyses, 
Earth Planet. Sci. Lett., 19, 290, 1973. 

Philpotts, J. A., C. C. Schnetzler, and 
S. R. Hart, Submarine basalts: Some 
K, Rb, Sr, Ba, rare-earth H 2 and C0 2 
data bearing on their alteration, modi- 
fication by plagioclase, and possible 
source materials, Earth Planet. Sci. 
Lett., 7, 293, 1969. 

Schilling, J. G., Iceland mantle plume: 


geochemical study of the Reykjanes lavas and MOR lavas. Although several 

Ridge, Nature, 21+2, 565, 1973. recent papers have dealt with the tec- 

Tatsumoto, M., C. E. Hedge, and A. E. tonic evolution of the Azores (Krause 

J. Engel, Potassium, rubidium, stron- and Watkins, 1970; Machado et al, 

tium, thorium, uranium, and the ratio 19 ?2; McKenzie, 1972), the evolution of 

of strontium-87 to strontium-86 in the region is still poorly understood, 

oceanic tholeiitic basalt, Science, 150, There is a moderate amount of petro- 

886 1965. logical data from the islands, but trace 

Thompson, G. A geochemical study of element and isotope data are scarce, 

the low-temperature interaction of Schilling (1975) has shown that ba- 

sea-water and oceanic igneous rocks, salts eru P ted alon S the Mid-Atlantic 

Eos 5U 1015 1973 ridge (MAR) transect of the Azores 

platform have anomalously high (La/ 

Sm) E . P . ratios and that there is a grada- 

Geochemistry of the Azores tion from high (La/Sm) E .F. values on 

and the Mid-Atlantic the Azores platform (39°N) to (La/ 

Ridge: 29°N to 60°N Sm) EF . values typical of MORB at 

Tirw/- n# TI7Z, •. c, 7 d tt . about 33°N. He has interpreted this as 

William M. White, Stanley R. Hart, . . ;1 .. „ ,11 ,~\r 

and Jean-Guy Schilling bem g the result of a mantle plume (Mor- 
gan, 1971) rising beneath the Azores and 

Introduction mixing southward along the ridge with 

Mid-ocean ridge basalts (MORB) are the low- velocity layer source of normal 

remarkable for their low abundances of no -g e basalts. 

large ion lithophile elements (LIL) such We re P ort here the results of a Sr- 

as K, Rb, Cs, Sr, Ba, and REE, and lsoto P e and alkah and , alkaline earth 

their low isotopic ratios of Sr and Pb, as trace elemen * ^ of J T avas *™ pted 

well as their general petrochemical uni- alon g the MAR from 2 ? N *° 6 ° N > on 

formity. The low concentrations of LIL, the Azores Islands and m the Terceira 

as well as other "incompatible" elements trou S h - The stud ^ d °™ments geochemi- 

and the low large-ion/small-ion ratios, cal sanation in the region and provides 

such as Rb/K and La/Sm in MORB, a test of hypotheses of Azores evolution, 

are generally thought to be due to their Particularly the plume hypothesis We 

derivation from a mantle severely de- also examine the relationship of tholeiitic 

pleted in these elements by previous basalt erupted on the MAR to the alkali 

episodes of melting (Gast, 1968). The lavas erupted on the Azores Islands, 

depleted character of MORB — which „ 7 . , . 7 L . 7 „ 7 

are, with very few exceptions, tholeiites Sampling and Analytical Procedure 

— contrasts strongly with that of tho- We analyzed 77 basalt samples from 

leiitic and alkali basalts erupted on 63 dredge stations along the MAR from 

oceanic islands. As both mid-ocean ridge 29°N to 60°N, shown in Fig. 56, and 41 

(MOR) and oceanic island volcanics are samples from the Azores Islands and 

mantle derived, this contrast implies the Terceira trough, shown in Figs. 57 

significant mantle heterogeneity and dif- and 58. Thirty-nine stations along the 

ferences in mantle evolution which war- ridge and two stations in the Terceira 

rant further investigation into their trough were dredged by R. V. Trident of 

nature. the University of Rhode Island during 

The Azores archipelago, which is the cruises TR89, TR100, TR119, TR122, 

locus of the triple junction of the North TR123, and TR154. Samples from other 

American, Eurasian, and African plates, stations were provided by several other 

is an excellent example of the contrast- investigators. Island samples were col- 

ing petrochemistry of oceanic island lected by two of us (White and Schill- 



E -A-F.Z, 

25° 20° W 

Fig. 56. Map of the study area. Solid circles designate R. V. Trident dredge stations; open 
circles designate other dredge stations. 

ing) and by M. D. Tapia. Nearly all the 
samples analyzed from the ridge are 
tholeiitic basalts. A few transitional and 
mildly alkaline basalts occur at 35° and 
45°N (Aumento, 1968; Muir and Tilley, 
1964; Schilling, 1975, and unpublished 
analyses), while lavas from the Azores 
Islands are alkalic. 

87 Sr/ 86 Sr ratios were measured using 
DTM's automated 6-inch mass spec- 
trometer. K, Rb, Cs, Ba, and in seven 
cases, Sr were analyzed by isotope dilu- 
tion. Details of the procedure and pre- 
cision of the isotope dilution analysis 
have been given elsewhere (Hart and 
Brooks, 1974). The remaining Sr analy- 
ses were done by x-ray fluorescence at 
the Geophysical Laboratory (precision: 
±3% at 100 ppm, 2 sigma) . 


Geochemical variation along the ridge. 
The 87 Sr/ 86 Sr profile of basalts erupted 
along the ridge is shown in Fig. 59. Two 
distinct maxima are apparent, one at 
44°-45°N, the other over the Azores 
platform (39°N). The maximum over 
the Azores platform (about 0.70345) 
appears slightly higher than that at 
45°N (about 0.70340). The transition 
zones north of 45° and south of 39°N are 
relatively gentle, with the 87 Sr/ 86 Sr 
ratios decreasing to typical ridge values 
of 0.70250 to 0.70270. Ten analyzed 
samples from the FAMOUS (French 
American Mid Ocean Undersea Studies) 
region, situated in the geochemical tran- 
sition zone south of the Azores, have in- 
termediate values. 




24° W 

Fig. 57. Map of the Azores Islands. Triangles indicate location of R. V. Trident dredge 
stations. 1000-meter and 3000-meter bathymetric contours are shown. 



5 Km 


Fig. 58. Sample location maps for Azores Islands. Solid circles designate sample locations. 
Dotted areas are historical lava flows. 






\ 0.7030 h 

00 0.7028 - 
0.7026 - 


2000 1500 1000 500 500 1000 

T-p 2 | | I I I I | I I I I | l ! I I | I I I I | I ! ! I | I 





/ \ 

/ 1 \ 

o • 





> > 











50 ( 

45° 40° 



Fig. 59. Variation of 87 Sr/ S6 Sr vs. radial distance from the center of the Azores 
(39°N, 30° W). Open symbols indicate values suspected of having changed due 
eruptional alteration. (2) indicates two or more identical analyses. (3) indicates three 


to post- 

Figure 60 shows the variation of K 
and Ba along the ridge; profiles of Rb, 
Cs, and Sr are similar. A comparison of 
Figs. 59 and 60 shows that the variation 
of the concentrations of these elements 
is quite similar to the variation of 87 Sr/ 
86 Sr though some differences exist. Trace 
element concentrations in basalts from 
the Azores platform are much higher 
than in those from the 45 °N area, 
whereas the 87 Sr/ 86 Sr values from the 
two areas are quite similar. The "sad- 
dle" in the 40°-43°N area dips rela- 
tively lower for the trace elements than 
for the 87 Sr/ 86 Sr ratios. Basalts from the 
Gibbs Fracture Zone tend to be some- 
what high in their trace element content. 

The variations of Ba/K and Rb/K 
along the ridge are similar; the former 
is shown in Fig. 61. While this profile 
exhibits a broad maximum over the 
Azores region, there are striking con- 
trasts between the variation of Ba/K 
and 87 Sr/ 86 Sr, notably the lack of a 
saddle between the Azores and 45°N. In 

addition, while the decrease of 87 Sr/ 86 Sr 
and trace element concentrations along 
the ridge south of the Azores is linear or 
slightly concave upward, the decrease of 
Ba/K is distinctly convex upward. The 
decrease of (La/Sm) E . F . is more or less 
linear (Schilling, 1975). 

87 Sr/ 86 Sr values for samples from the 
35°N Oceanographer Fracture Zone area 
are anomalously high. In basalts from 
stations TR119 8D and 7D this anoma- 
lous character includes trace element 
and major element chemistry as well. 
However, these basalts are vesicular 
(20% vesicles) with sediment in the 
vesicles, and high 87 Sr/ 8r, Sr values could 
be due to sediment contamination. Simi- 
larly, basalts from station V30 RD7 
(one analysis) show evidence in thin 
section of alteration and have somewhat 
high H 2 contents. Finally, 87 Sr/ 86 Sr 
values for station TR123 4D (33°N) 
are also anomalously high. These sam- 
ples are quite unusual in composition, 
with extremely low LIL-element con- 








2000 1500 1000 500 500 1000 1500 

TTT& I | I I I I | I \f\ I | I I I I | I I I I | I 









<•• • 


• • ••• 

• • 






• • 

• #f * • 








45° 40° 




2000 1500 1000 500 500 1000 1500 

i I I | & I | I I I I | I I I I | I I l I | I I I I | I I I I | I I I I | I I 









• ** • 

• • 

# • • 



• (2) 

I 9 


• UJ 


^ < 

• • *s ••• %1 T^ 





(2) . * (2 ) % \ 


45° 40° 


35 « 

30° N 

Fig. 60. a, variation of K (ppm) vs. radial distance from the center of the Azores platform. 
b, variation of Ba (ppm) vs. radial distance from the center of the Azores platform. Symbols 
are the same as for Fig. 59. 




' ' 1 







2000 1500 1000 500 500 1000 1500 

yll i | i i i i | i i i i | i i i i | i I i i | i i I l | i i I 1 | i I 




• • 



o • 



•• *• L. • • 


• ••• 

• •» ~* L» • w • 

•IT* • 

• •*. • • ' #(2) . 

• | 

T 2 )% •" i * 





45° 40° 


35 ' 

30° N 

Fig. 61. Variation of Ba/K vs. radial distance from the center of the Azores platform. Sym- 
bols are the same as for Fig. 59. 

tents and low trace element ratios such 
as La/Sm and Rb/K. They tend to be 
picritic, and relatively Al rich. 

Basalts from the region of 50°N have 
87 Sr/ 86 Sr values (0.70233) among the 
lowest yet obtained from MOR, while 
their trace element chemistry is very 
close to the average for MORB. These 
low values are an important feature of 
the profile, although their significance is 
not yet clear. 

Azores Islands and the Terceira 
trough. The samples from the Azores 
Islands and the Terceira trough vary 
from alkali basalts to highly differen- 
tiated commendites. The 87 Sr/ 86 Sr ratios 
in these lavas show wide variation: 
0.70332 to 0.70525; however, all but six 
values fall in the range 0.70332 to 
0.70394, and over half the samples stud- 
ied have values less than 0.70354. Our 
discussion of Azores Island geochemistry 
will center on 87 Sr/ 86 Sr results, as major 
element analysis necessary for the inter- 
pretation of the trace element data is 

not yet available for all the samples. 
The isotope data are plotted against 
longitude in Fig. 62. 

Flores and Corvo, both volcanically 
extinct, lie along a N-S trending sub- 
marine ridge and are the only two is- 
lands west of the MAR. 87 Sr/ 86 Sr values 
from these two islands range from 
0.70332 to 0.70352, with no apparent 
isotopic differences between the two 
islands. Faial and Pico, both of which 
have had several historical eruptions, are 
aligned along a submarine ridge, which 
apparently extends to Sao Miguel 
(Krause and MacGregor, 1975). 87 Sr/ 
8G Sr ratios from these islands range from 
0.70347 to 0.70394, the values from Faial 
tending to average somewhat higher than 
those from Pico. Sao Jorge, a young, ac- 
tive, linear volcano lies along a sub- 
marine ridge which also apparently 
extends to Sao Miguel (Krause and 
MacGregor, 1975) , and has 87 Sr/ 86 Sr in 
the range 0.70332-0.70338. 

Graciosa, Terceira, and Sao Miguel lie 






0.704 - 


00 0.703 












\ U D J»v 









































- 1 O <*■ 

< ° O 
— O ~ 

< — ^ 







Li_ 0- O) 
1 1 






30 c 

2 9 c 


2 7< 

2 6 c 



Fig. 62. Variation of 87 Sr/ 86 Sr vs. longitude across the Azores plateau. Triangles denote 
dredge stations; squares denote samples from Faial and Pico; hexagons denote samples from 
Sao Jorge; circles denote samples from Flores, Corvo, Santa Maria, Graciosa, Terceira, and 
Sao Miguel. Open symbols are basalts and trachybasalts ; closed symbols are trachytes. Solid 
line encloses samples from the MAR and the Treciera line ; dotted line encloses samples from 
Sao Jorge. Dashed line encloses samples from Faial and Pico. 

along the Terceira trough, which may be 
a secondary spreading center (Krause 
and Watkins, 1970). There have been 
several historic eruptions on Sao Miguel 
and Terceira, and numerous submarine 
eruptions reported from the trough; al- 
though there have been no eruptions on 
Graciosa, evidence suggests it is still 
active (Machado, 1967). 87 Sr/ 86 Sr ratios 
from Graciosa, Terciera, and submarine 
basalts from the trough fall in the range 
0.70336-0.70345, with the exception of 
two commendites from Graciosa and 
Terciera, which have higher ratios. 

The high Rb/Sr ratios (around 10) in 

these two commendites suggest that the 
87 Sr/ 86 Sr ratios are the result of 87 Sr 
accumulation due to Rb decay since dif- 
ferentiation of the magma. If the initial 
ratios of these two samples were near 
the average for the two islands (about 
0.70350), then the apparent ages are 2.5 
m.y. for the Graciosa commendite and 
0.75 m.y. for the Terceira sample. Field 
relationships suggest these ages are rea- 

Sao Miguel, which is apparently the 
intersection of the Terceira, Sao Jorge, 
and Faial-Pico tectonic features, is geo- 
chemically anomalous. One basalt from 



the western third of the island (trend- 
ing NW-SE) has an 87 Sr/ 8(J Sr ratio of 
0.70337, while four values from the east- 
ern two-thirds of the island (trending 
E-W) range from 0.70434 to 0.70525. 

Samples from Santa Maria, which is 
not aligned along any tectonic feature 
and which clearly is the oldest of the 
islands, yielded 87 Sr/ 80 Sr ratios of 
0.70336 to 0.70352. 

Discussion and Conclusions 
Comparison with Iceland. Maximal 

values for 87 Sr/ 86 Sr and (La/Sm) E . F ., 
0.70345 and 2.8, respectively, from the 
ridge transect of the Azores platform are 
considerably higher than values from 
the ridge transect across Iceland, 0.70305 
and 1.3 for 87 Sr/ 86 Sr and (La/Sm) E . F ., re- 
spectively (Hart et al, 1973). 87 Sr/ 86 Sr 
and (La/Sm) E .F. values from 45°N are 
slightly lower than those from the Azores 
platform. In addition, while the Iceland- 
Reykjanes Ridge transition zone be- 
tween the undepleted Icelandic basalts 
and the depleted ridge basalts extends 
only some 400 km, the transition zone 
between the Azores platform and the 
" normal" ridge to the south extends 
some 1000 km. Thus the geochemical 
anomaly beneath the Azores is greater 
in both magnitude and extent. 

The marked isotopic uniformity along 
the central neovolcanic zones of Iceland 
(Hart et al., 1973) is not duplicated in 
the MAR transect of the Azores plat- 
form, as the 87 Sr/ 8(5 Sr values begin de- 
creasing within 100 km north and south 
of the center of the platform. 

Variation patterns in the Azores. Ini- 
tial 87 Sr/ 86 Sr ratios from six of the 
islands fall in the narrow range of 
0.70332 to 0.70354, similar to— though 
on the average slightly higher than — 
87 Sr/ 86 Sr ratios of basalts from the MAR 
transect of the platform. 87 Sr/ 86 Sr ratios 
tend to conform to the geographical dis- 
tribution of the islands; that is, islands 
lying along the same tectonic feature 
tend to have similar 87 Sr/ 86 Sr ratios with 
the obvious exception of eastern Sao 

Miguel. As Fig. 62 indicates, there is no 
systematic change in 87 Sr/ 8(J Sr ratios 
longitudinally away from the ridge. 

Faial, Pico, and particularly Sao 
Miguel are the only islands where re- 
liable 87 Sr/ 86 Sr ratios are markedly dis- 
similar from those of the adjacent MAR. 
All the samples are fresh and the high 
87 Sr/ 86 Sr ratios cannot be explained by 
post-eruptional alteration. The situation 
is somewhat analogous to that of the 
Snaefells Peninsula in Iceland, where 
lavas are erupted with distinctly higher 
87 Sr/ 86 Sr ratios (0.70325) than those of 
the central neovolcanic zones (0.70304) 
(Hart et al., 1973). However, the magni- 
tude of the difference is considerably 
greater in the present case. Why lavas 
with such greatly differing isotopic ra- 
tios can be erupted on a single island 
while variation along the ridge is so 
gradual over distances up to 1000 km, 
remains an enigma. 

Mantle geochemistry and the plume 
hypothesis. A clear analogy exists be- 
tween the results presented here and 
those from the Reykjanes Ridge (Hart 
et al., 1973). Schilling (1973) has inter- 
preted the geochemical variation along 
the Reykjanes Ridge as being the result 
of the upwelling of a mantle plume be- 
neath Iceland and its mixing with the 
depleted mantle source of MAR basalts. 
Additional geophysical and geochemical 
evidence supports this interpretation 
(Francis, 1973; Vogt and Johnson, 1973; 
Hart et al, 1973). O'Hara (1973, 1975) 
has challenged this interpretation and 
has suggested that isotope ratios such 
as 87 Sr/ 86 Sr and 2O(5 Pb/ 204 Pb may be 
changed somehow during fractional 
crystallization, but he has not suggested 
a mechanism to produce such changes. 
The island isotope data presented here 
show no systematic variation with dif- 
ferentiation, and, as there is no evidence 
in the literature to support O'Hara's 
suggestion, we consider it improbable 
that the Sr isotope ratios can be affected 
by fractional crystallization. 

Recently it has been suggested that 



isotopic equilibrium may not be achieved 
during the melting process and that this 
disequilibrium may explain the observed 
variation in 87 Sr/ 86 Sr along the Reyk- 
janes Ridge (Sigvaldason et al., 1974; 
O'Nions and Pankhurst, 1974). O'Nions 
and Pankhurst (1974) have presented a 
model for disequilibrium melting of a 
phlogopite-bearing mantle. Their model 
predicts that for small degrees of melt- 
ing, the 87 Sr/ 86 Sr ratio of the melt would 
be significantly higher than that of the 
source, and that the 87 Sr/ 86 Sr ratio of 
the melt would decrease with increasing 
degree of melting. Thus, according to 
this model, liquids produced by small 
degrees of melting, such as nephelinites 
(Green, 1970) and alkali basalts, should 
have higher 87 Sr/ 86 Sr ratios than tho- 
leiites produced by larger degrees of 
melting by the same source. But on 
Oahu, nephelinites have ratios similar to 
or slightly lower than tholeiites (Powell 
and Delong, 1966). In the present case, 
most alkali lavas from the Azores Is- 
lands have 87 Sr/ 86 Sr ratios similar to 
those of tholeiites from the MAR tran- 
sect of the Azores platform. Further- 
more, as U and Th are not concentrated 
in phlogopite, disequilibrium melting of 
a phlogopite-bearing mantle cannot ex- 
plain the variation of lead isotopes along 
the Reykjanes Ridge (Sun et al, 1975) 
or the observed enrichment in radio- 
genic Pb in Azores Islands lavas (Sun 
and Hanson, 1975; Oversby and Gast, 
1968). Thus, while we agree that the 
possibility of disequilibrium melting 
needs further investigation, the model 
presented by O'Nions and Pankhurst 
(1974) does not satisfactorily explain 
the observed variation of radiogenic 
isotopes in oceanic lavas. For further 
discussion of the disequilibrium melting 
hypothesis, see the accompanying paper 
(this Report) by Hofmann and Hart. 

Therefore, we tentatively conclude 
that the observed geochemical variation 
along the Ridge is due to geochemical 
variation in the mantle source. The ob- 
served variation along the Ridge is con- 

sistent with the Azores mantle plume 
hypothesis as considered by Schilling 
(1975), although geochemical studies 
alone cannot prove the existence of 
plumes, which are essentially dynamical 
concepts. The data would fit alternative 
models such as a vertically or horizon- 
tally inhomogenous static mantle. How- 
ever, static mantle models have diffi- 
culty explaining how the mantle under- 
lying spreading ridges can produce ba- 
saltic magmas of uniform composition, 
apparently indefinitely. Static models 
also have difficulty explaining the grad- 
ual transitions observed along the MAR 
and Reykjanes Ridge. 

It is unclear whether the 45°N area 
represents a separate anomalous mantle, 
or is in some way associated with the 
anomalous mantle beneath the Azores, 
or whether the isolated anomalous re- 
gions at 35°N and 33°N can be ac- 
counted for by the plume hypothesis. 

If the plume interpretation is correct, 
then our results indicate that there is no 
diminution of the plume's geochemical 
effect longitudinally as far as Sao Mi- 
guel and Santa Maria. Its effect there- 
fore is not radially symmetrical as 
87 Sr/ 86 Sr ratios begin decreasing along 
the MAR within 100 to 200 km from the 
center of the platform. Our analyses in- 
clude both historical flows and flows 
from some of the oldest volcanic se- 
quences of the archipelago (Santa Maria 
and NE Sao Miguel; Abdel-Monem 
et al., 1968) and reveal no particular 
systematic change with time. Hence, the 
plume must have been present beneath 
the Azores for at least the last 4 to 6 

Lavas from 6 of the Azores Islands all 
have initial 87 Sr/ 86 Sr ratios which fall 
within the range of 87 Sr/ 86 Sr ratios in 
basalts from the MAR transect of the 
Azores platform. This implies that alkali 
basalts on these islands are being de- 
rived from a mantle geochemically simi- 
lar to that which is feeding volcanism 
along the MAR transect of the platform. 
However, higher values on Faial, Pico, 



and Sao Miguel imply considerable local 
mantle heterogeneity in this region. 

At present we have no adequate ex- 
planation for the differences between 
LIL, LIL ratio, and 87 Sr/ 86 Sr profiles 
presented here. However, our investiga- 
tion into these differences centers on 
these possibilities: variable degree of 
partial melting and/or crystal fractiona- 
tion, variation in source mineralogy, 
second-order variation in mantle geo- 
chemistry, and mixing models which 
might account for the observed differ- 
ences. These differences may be impor- 
tant clues to magmatic processes and 
mantle dynamics in this region. 

This work represents a part of the 
doctoral dissertation of one of us (Wil- 
liam White) at the University of Rhode 
Island and much of the discussion pre- 
sented here is preliminary in nature. 


Abdel-Monem, A., L. A. Fernandez, and 
G. A. Boone, Pliocene-Pleistocene 
minimum K-Ar ages of older eruptive 
centers, Eastern Azores (abstr.), 
Trans. Am. Geophys. Union, 1^9, 363, 

Aumento, F., The mid-Atlantic Ridge 
near 45°N. II. Basalts from the area 
of Confederation Peak, Can. J. Earth 
Sci., 5, 1-21, 1968. 

Francis, T. J. G., The seismicity of the 
Reykjanes Ridge, Earth Planet. Sci. 
Lett, 18, 119-124, 1973. 

Gast, P. W., Trace element fractiona- 
tion and the origin of tholeiitic and 
alkaline magma types, Geochim. Cos- 
mochim. Acta., 32, 1057-1086, 1968. 

Green, D. H., The origin of basaltic and 
nephelinitic magmas, Trans. Leiches- 
ter Liter. Philos. Soc, 64, 26-54, 1970. 

Hart, S. R., and C. Brooks, Clinopyrox- 
ene matrix partioning of K, Rb, Cs, 
Sr, and Ba, Geochim. Cosmochim. 
Acta, 28, 1799-1806, 1974. 

Hart, S. R., J-G. Schilling, and J. L. 
Powell, Basalts from Iceland and 
along the Reykjanes Ridge; Sr isotope 

geochemistry, Nature, Phys. Sci., 246, 
104-107, 1973. 

Krause, D. C., and B. A. McGregor, 
Azores plateau submarine geology, 
Geol. Soc. Am. Bull., in press, 1975. 

Krause, D. C, and N. D. Watkins, 
North Atlantic crustal genesis in the 
vicinity of the Azores, Geophys. J. R. 
Astron. Soc, 19, 261-283, 1970. 

Machado, F., J. Quintino, and J. H. 
Monteiro, Geology of the Azores and 
the mid-Atlantic Rift, Twenty-fourth 
Int. Geol. Congr., Sec. 3, pp. 134-142, 

McKenzie, D. P., Active tectonics of the 
Mediterranean region, Geophys. J. R. 
Astron. Soc, 30, 109-185, 1972. 

Morgan, W. J., Convection plumes in the 
lower mantle, Nature, 230, 42-43, 

Muir, I. D., and C. E. Tilley, Basalts 
from the northern part of the rift zone 
of the mid-Atlantic Ridge, J. Petrol., 
5, 409-434, 1964. 

O'Hara, M. J., Non-primary magmas 
and dubious mantle plume beneath 
Iceland, Nature, 243, 507-508, 1973. 

O'Hara, M. J., Is there an Icelandic 
mantle plume? Nature, 253, 708-710, 

O'Nions, R. K., and R. J. Pankhurst, 
Petrogenetic significance of isotope 
and trace element variation in vol- 
canic rocks from the mid-Atlantic, J. 
Petrol, 15, 603-634, 1974. 

Oversby, V. M., and P. W. Gast, Lead 
isotope compositions and uranium de- 
cay series disequilibrium in recent 
volcanic rocks, Earth Planet. Sci. 
Lett., 5, 199-206,1968. 

Powell, J. L., and S. E. Delong, Isotopic 
composition of strontium in volcanic 
rocks from Oahu, Science, 153, 1239- 
1242, 1966. 

Schilling, J-G., Iceland mantle plume: 
Geochemical study of the Reykjanes 
Ridge, Nature, 242, 565-571, 1973. 

Schilling, J-G., Azores mantle blob: rare 
earth evidence, Earth Planet. Sci. 
Lett, 25, 103-115, 1975. 



Sigvaldason, G. E., S. Steinthorsson, N. 
Oskarsson, and P. Imsland, Composi- 
tional variation in recent Icelandic 
tholeiites and the Kverkfjoll hot spot, 
Nature, 251, 579-582, 1974. 

Sun, S. S., and G. N. Hanson, Evolution 
of the mantle: geochemical evidence 
from alkali basalts, Geology, 3, 297- 
302, 1975. 

Sun, S. S., M. Tatsumoto, and J-G. 
Schilling, Mantle mixing along the 
Reykjanes Ridge axis: lead isotopic 
evidence, submitted to Science, 1975. 

Vogt, P. R., and G. L. Johnson, A longi- 
tudinal seismic reflection profile of 
Reykjanes Ridge: part II — implica- 
tions for the mantle hotspot hypothe- 
sis, Earth Planet. Sci. Lett., 18, 49- 
58, 1973. 

Secondary Minerals in 
Oceanic Basalts 

Manuel N . Bass 

The results of a study of the secondary 
minerals in basalts cored during Leg 34 
of the Deep Sea Drilling Project were 
combined with earlier results on basalts 
from Leg 17 (Bass et al, 1973a, 19736) 
and the many extant observations on 
dredged basalts to construct a composite 
picture of the progressive development 
of secondary minerals in oceanic basalts 
during late magmatic and deuteric proc- 
esses, and during subsequent low-tem- 
perature processes when the rocks were 
exposed to open, oxygenated bottom 
waters (the so-called "seawater altera- 
tion") and to pore waters, the oxygen 
contents of which varied with time 

For mnemonic purposes, we can dis- 
tinguish four groups of processes: (1) 
late magmatic-deuteric ; (2) sea-water 
alteration (pre-burial) ; (3) nonoxidative 
diagenesis (post-burial, limited oxida- 
tion) ; and (4) oxidative diagenesis 
(post-burial, intense oxidation). 

These four discrete stages are an over- 
simplification since, prior to burial, 
stages 2, 3, and 4 proceed simultaneously. 

The interior portions of unfractured 
blocks, whose exteriors are bathed in 
open bottom water and are undergoing 
sea-water alteration in a direct sense, 
are subject at successively inward levels, 
to stages 4 (oxidation superimposed on 
earlier hydration) and 3 (hydration with 
limited oxidation). The successive levels, 
which in part control the color zonation 
seen in joint blocks of oceanic basalt, 
reflect the fact that water is available in 
abundance and diffuses relatively fast, 
whereas oxygen supply is limited and its 
inward progress is retarded either by a 
slow diffusion rate, or, more likely, by its 
consumption by readily oxidized primary 
and secondary phases. The degree to 
which the concurrent processes penetrate 
into a given igneous body or pile of flows 
depends on the porosity and permeabil- 
ity of the system. At an actively spread- 
ing ridge crest these are controlled by 
pillow structure, vesicles and diktytaxitic 
vugs, primary cooling cracks and tec- 
tonic effects; and the alteration zones 
penetrate into essentially fresh rock. In 
old crust reexposed on tectonic ridges 
and in troughs of fracture zones, the 
early porosity and permeability have 
been decreased by deposition of veins 
and amygdules during burial, and in- 
creased by tectonic fracturing during 
burial and especially during the subse- 
quent activity that led to reexposure of 
the rocks. Processes 3 and 4 continued 
during the burial of the rocks, as dis- 
cussed more -fully below. Following re- 
exposure, process 2 is renewed on exposed 
surfaces and a new wave of concurrently 
developing zones sweeps through the old 
rocks, superimposing themselves on any 
earlier alteration zones developed during 
the earlier history of the rocks. It is this 
type of complex history that is recorded 
in many or most old rocks dredged from 
fracture zones. For the remainder of this 
discussion we ignore the multi-stage 
complications consequent on reexposure 
after a long history of burial and focus 
on the alteration of rocks formed at the 
ridge crest and subsequently buried, but 


not reexposed, and of sills injected into biguous cases in which to study the suc- 
such rocks after burial. cessive effects of nonoxidative and oxi- 
Following burial, process 2 (seawater dative diagenesis, namely those resulting 
alteration) is terminated, but processes from processes 3 and 4. I had earlier 
3 and 4 proceed as before, penetrating studied samples from sills cored during 
natural joint blocks in successive waves Leg 17 (hole 169 in particular, and pos- 
recorded as visible color zones which sibly holes 167 and 170 as well). I was 
reflect various secondary minerals. The especially fortunate to participate in Leg 
reducing capacity of the system is low, 34 during which, for the first time, cores 
and redox potentials relatively high, so of massive flows of mid-ocean ridge 
oxidation effects produced during earlier basalts (MORB) were recovered, yield- 
stages, such as sea-water alteration of ing the best cores from that leg. The old 
exposed surfaces, are not reversed after (Cretaceous) sills from Leg 17 and the 
burial. In other words, the rocks seem, relatively young (Miocene to late Eo- 
for the most part, always to retain the cene) massive flows from Leg 34 comple- 
most oxidized state they have attained, ment each other in age, thus allowing 
Near an active ridge crest, joints are the composite picture to be drawn in 
numerous and open, and pore fluids pene- considerable detail. 

trate easily to the surfaces of the joint Before outlining that picture, however, 
blocks. After burial the development of one additional factor needs to be dis- 
veins and amygdules is accelerated, and cussed, namely susceptibility to altera- 
fracture porosity tends to decrease. Near tion of the various primary and see- 
the ridge crest the tectonic activity con- ondary phases. This affects the rate of 
nected with the spreading process creates alteration, the rate of development of 
new fractures and tends to counteract distinct alteration zones, and the clarity 
the sealing effect of veins. Away from of those zones. The main phases in vir- 
tue crest the tectonic activity diminishes ; tually all noncrystalline oceanic basalts 
vein deposition dominates; porosity is are clinopyroxene (mainly augite) and 
sealed; fractures are healed; and thereby plagioclase. Subordinate phases are ti- 
the bulk density and elasticity of oceanic tanomagnetite, olivine, and interstitial 
crustal layers 2 and 3 increase. The rate glass. Volumetrically the chilled glass 
of further alteration then tends to be surfaces of flows and pillows are minor, 
dominated more and more by diffusion, but they assume great significance be- 
and successive steps in the alteration cause of the frequency with which they 
process are more faithfully recorded as are recovered in cores and dredges, and 
visible zones or replacement effects their unique importance in marking sur- 
(pseudomorphs) . In a highly fractured faces of discrete igneous cooling units, be 
system one would have to study many they pillow, flows, or sills. Of these vari- 
joint blocks to reconstruct the sequence ous phases, clinopyroxene and plagioclase 
of events. In the nonpillowed, less frac- are most resistant to seawater alteration 
tured, massive portions of flows, and and diagenesis. Only in rocks as old as 
in sills injected far from ridge crests, Cretaceous does one begin to see altera- 
that are thereby not subject to severe tion of plagioclase, either to an alumi- 
tectonism, diffusion dominates from the nous smectite, seen only along cracks or 
start and the diffusion fronts tend to in selected zones, or to a monoclinic K- 
parallel the generally flat-lying flow or feldspar pseudomorph, the extent of 
sill boundaries. Thus, Deep Sea Drilling development of which depends on the 
Project (DSDP) holes tend to penetrate supply of potassium, which in turn de- 
the alteration zones at a high angle and pends on the abundance of smectite 
core recovery is best in such unfractured which rapidly takes up K in exchange 
rocks. These cores provide the least am- sites and must be virtually saturated 



(i.e., at equilibrium with pore waters) 
before K is available to partake in re- 
placement of plagioclase. Clinopyroxene 
likewise shows almost no alteration in 
rocks younger than Cretaceous. In them 
it is replaced under nonoxidative condi- 
tions by smectite. The alteration is 
greatly accelerated by high activities of 
carbonate ion species, so that highly al- 
tered clinopyroxene is generally associ- 
ated with abundant carbonate. By the 
time clinopyroxene and plagioclase begin 
to show alteration, all other primary 
phases have been extensively altered, and 
the progress of that alteration can be 
followed in the relatively young rocks 
cored during Leg 34. Titanomagnetite is 
oxidized at high temperatures to Ti-poor 
magnetite and ilmenite, and at low tem- 
peratures to titaniferous maghemomag- 
netite and maghemite. Olivine is replaced 
by smectite under nonoxidative condi- 
tions, and is either wholly removed or 
replaced by porous, earthy Fe oxide 
under oxidative conditions. Smectite rims 
around earthy Fe oxide cores in pseudo- 
morphs after olivine record a composite 
history. Fe-rich olivine, which is re- 
stricted to the matrices of noncrystalline 
MORB, alters much more rapidly than 
early, Mg-rich olivine. Hence, the ex- 
istence of groundmass olivine in MORB, 
though permitted by known phase rela- 
tions, was not firmly established until 
rare remnants and much more abundant 
pseudomorphs after it were found in the 
massive, diabasic-textured MORB cored 
during Leg 34 at sites 319 and 321. Inter- 
stitial glass is presumably richer in Fe 
and water in coarser grained, more 
slowly cooled rocks, and hence, is more 
rapidly altered than is the relatively 
Mg-rich glass of chilled selvages. The 
interstitial glass is altered to a green or 
blue-green, relatively Fe-rich smectite, 
and alters even more rapidly than associ- 
ated olivine if one may judge from the 
relative frequency and abundance of 
fresh remnants of these phases. The 
glassy, relatively Mg-rich selvages of 
pillows and flows alter progressively 

through golden yellow to orange iso- 
tropic palagonite; to red isotropic, highly 
oxidized palagonite; or red-to-orange 
fibropalagonite (a dense aggregate of 
poorly crystallized smectite crystallites). 

Late magmatic-deuteric products. Late 
hydrous products of the magmatic stage 
include rare biotite (and possibly amphi- 
bole). Quench products include inter- 
stitial hydrous glass and possibly quench 
smectite. Deuteric products include Ti- 
poor magnetite and maghemomagnetite 
with ilmenite lamellae pseudomorphic 
after titanomagnetite; paper-thin veins 
of chlorite and minor talc; rare idding- 
site; and possibly smectite after olivine 
and interstitial glass. Minor talc may 
accompany smectite after olivine. 

Products of sea-water alteration. Pa- 
lagonite and fibropalagonite are the most 
prominent products of this stage. Other 
products formed concurrently just below 
or inward from the surfaces of joint 
blocks are most readily recognized in 
those blocks, one of the surfaces of which 
is a chilled glassy selvage. In and just 
inward from the glass selvage, olivine 
and interstitial glass are altered to 
earthy Fe oxide where they are exposed 
to abundant oxygen. Where access of 
oxygen is retarded and an early hydra- 
tion front has passed through, the olivine 
is altered to smectite and the interstitial 
glass to golden yellow palagonite or 
fibropalagonite. A later wave of oxida- 
tion may replace remnant olivine by Fe 
oxide to give composite pseudomorphs, 
and convert the altered glass from golden 
yellow to brown, orange, or red. Still 
further inward the interstitial glass is 
hydrated but not oxidized, and is re- 
placed by relatively well-crystallized 
green or blue-green smectite which is 
difficult to assign with certainty to the 
stage of seawater alteration (pre-burial) 
or diagenesis. 

Fe and Mn oxides formed during sea- 
water alteration are x-ray amorphous 
near the surfaces of joint blocks. As little 
as two of three centimeters inward, how- 
ever, dark veins give x-ray patterns of 


todorokite, birnessite (which may have Veins are composed mainly of green 
formed directly or replaced todorokite), to blue-green smectite, magnesian cal- 
and, rarely, goethite. The crystallinity cite, aragonite, and a mixture of pyrite 
seems to be a function of rapidity of crys- and marcasite. Lesser constituents are 
talization, hence of rates of nucleation celadonite, chlorite, talc, and mixed- 
and crystal growth. Phillipsite veins are layer clay minerals. The chemical con- 
frequently found in those rocks which re- stituents of the smectite are probably 
cord seawater alteration, but, except derived from the immediate wall rocks 
well inward from surfaces, the veins are since smectite veins seem to die out 
believed to have formed after burial individually and collectively toward 
since phillipsite appears to be unstable contacts of cooling units. The relatively 
in open seawater. Carbonate is minor or high ferric Fe content (Table 9) suggests 
rare in the phillipsite and Mn oxide some oxidation, but limited in extent, so 
veins and appears to have formed pri- that the smectite lattice accommodates 
marily after burial. the ferric Fe and independent ferric Fe 
Products of nonoxidative diagenesis. phases do not form. The crystallization 
Interstitial glass is replaced by green to of smectite keeps the activity of Mg ions 
blue-green smectite (some of which may depressed to levels at which the nucle- 
be of quench origin). Groundmass oli- ation of calcite is not inhibited. The 
vine is wholly replaced, and phenocrystic textural and compositional relations of 
olivine is partially (along cracks and the carbonates suggest that aragonite 
edges) to wholly replaced by similar formed stably or metastably and later 
smectite. The high ferric Fe content of than magnesian calcite; for instance, the 
the blue-green smectite in veins associ- aragonite is clear, unetched, and often 
ated with the altered olivines and glass idioblastic, suggesting that its rate of 
(Table 9) suggests that limited oxida- inversion to calcite was effectively nil. 
tion occurs at this stage, but not enough From this fact it is inferred that as 
to produce independent ferric Fe phases, smectite crystallization slowed, the ac- 
Titanomagnetite may undergo some oxi- tivity of Mg ions in pore fluids rose, the 
dation toward maghemomagnetite. nucleation of calcite ceased, its growth 

TABLE 9. Wet Chemical Analysis of Green Vein Smectite 

Formed during Nonoxidative Diagenesis, Hole 321, 

Core 14, Section 3, 80-100 cm. 

Original (%) 

Corrected* (%) 

Si0 2 



Ti0 2 



A1 2 3 



Fe 2 3 















Na 2 



K 2 



H 2 + 



H 2 0- 



P 2 5 






CI 72 ppm 

* 2.37 Plagioclase (An 5B ) subtracted, normalized to 100% 
(H. Onuki, Analyst). 



slowed radically or stopped, and ulti- 
mately, aragonite formed. Concurrently 
a second generation of chlorite (pre- 
sumably Mg-rich) began to form. It is 
intimately intergrown with late, light- 
blue (presumably Mg-rich) smectite, 
which often occupies the centers of thick 
smectite veins. Thus, while oxygen, H 2 
and C0 2 appear to be added, little if 
any material appears to be lost, so that 
a vertical section through a cooling unit 
should yield the original, nonvolatile 
composition of the unit. Like smectite 

veins, the carbonate veins die out toward 
surfaces of cooling units. 

The origin of the sulfides is an un- 
solved puzzle in two ways. Biological 
reduction of pore-water sulfate is an in- 
adequate source of sulfur. The host rock 
is an adequate source, but the mecha- 
nism of its mobilization is unknown. The 
reason for the coexistence of two sulfides 
is likewise unknown, despite the fact 
that the coexistence is a long-known fact 
of many low-temperature sulfide occur- 
rences. In the Cretaceous rocks from Leg 

from d (211) 


Holes 319 319A 




Hole 321 



Hole 1.65 A 

o Associated with green smectite 
• Associated with aragonite 
^ In amygdules or irregular patches 

° At outer contacts of or between 
flows or pillows 

( J In veins 
p Microprobe analysis 



1 100 




Hole 169 

h-^Hole 171 (Top of Horizon 



Hole 166 


Hole 167 (Top of Magellan 

1 2 3 4 5 
Mole % MgC0 3 -> 


Fig. 63. Compositions of secondary calcites from basalts cored during legs 17 and 34 plotted 
against ages of basal sediments. The empty large squares refer to calcites associated with 
ferric oxides and/or oxidized smectites; note that they tend to be lower in Mg than those 
associated with aragonite and green smectite. 



17 no marcasite was found, but its pres- 
ence cannot be excluded since the Leg 
17 rocks are mostly oxidatively altered 
and little sulfide of any sort survives. 
Therefore, the presently known occur- 
rences of marcasite in oceanic crustal 
basalts of Miocene and late Eocene ages 
do not provide an adequate basis for 
speculations on its rate of inversion to 
pyrite at low temperatures. 

Mn oxides formed at this stage of 
alteration are mainly or solely todoro- 
kite, which is fairly well crystallized, 
presumably as a result of slow rate of 
crystallization. Phillipsite formation at 
this stage was restricted to chilled sur- 
faces. A possible reason is the inability 

of phillipsite to compete effectively with 
smectite, celadonite, and talc for the 
limited available dissolved silica in the 
interior of the cooling units, and its 
much more effective competition near 
glassy surfaces, the alteration of which 
provided relatively abundant free silica 
and few or no competing phyllosilicates. 
Products of oxidative diagenesis. The 
advent of oxygen behind sharp, diffusion- 
controlled oxidation fronts inaugurates 
oxidative diagenesis. Any remnant oli- 
vine is oxidized to earthy Fe oxide, and 
titanomagnetite to titaniferous mag- 
hemomagnetite and maghemite. The 
outermost, Fe-rich rims of clinopyroxene 
may be oxidized. Smectite is oxidized, 

Natural minerals : 

a Hole 321 

• Hole 58, with 0%&10°/oOf Fe 

as Fe ++ (atomic °/o) 
x Hole 250A 

"Representative" analysis 
Ross & Hendricks, 1941 

• Synthetic celadonite (Velde, 
1972) and ideal: 
KMgFe +++ Si 4 O 10 (OH) 2 


= AI + Fe +++ ) 

Goethite, Hematite, 



(Fe +Mg=)F 

Fig. 64. Modified AKF diagram (quartz generally absent) showing the diagenetic phyllo- 
silicate assemblages in oceanic crustal basalts. In some respects the diagram is an overlay of 
an ACF on an AKF diagram. The chlorite field is guesswork deployed for convenience of 
illustration. Calcite- and aragonite-bearing assemblages are not shown. The calcite and 
aragonite would plot at the apex K, which is also the apex C of the ACF overlay. They com- 
monly coexist with smectite and talc, less consistently with chlorite and celadonite. These 
assemblages and those observed to contain four phyllosilicates could be plotted only by 
separation of one of the composite components shown, in order to create a three-dimensional 
tetrahedral volume. This could be done without violation of phase-rule restrictions. The 
oxidation states of the phyllosilicates shown increase monotonically from the base toward 
the apex A. 



and charge balance is preserved by ex- 
pulsion of ferric Fe as a red, brown, or 
yellow dust which pigments the intrinsi- 
cally colorless resulting smectite. Except 
for some celadonite, the formation of 
which is limited by the supply of K, no 
new phyllosilicates form. Sulfides are 
oxidized, resulting in limonite stains in 
host rock adjacent to formerly sulfide- 
bearing veins. The limonite is the most 
readily visible sign of the oxidative al- 
teration. Aragonite inverts to Mg-poor 
calcite. Magnesian calcite adjusts pro- 
gressively toward stabler calcites poorer 
in Mg, a change which appears better 
correlated with oxidation than with the 
age of the host rock (Fig. 63). 

Phillipsite continues to form. If joints 
and veins are opened and admit Si-rich 
pore fluids from the surfaces of cooling 
units to the interiors, phillipsite and Mg- 
poor calcite may form drusy vugs in the 
interiors of cooling units where the cry- 
stallization may be slow enough to form 
well-crystallized birnessite, either by 
oxidation of earlier todorokite or by di- 
rect precipitation of birnessite. Fairly 
well-crystallized goethite may form un- 
der similar circumstances. 

The observed phase assemblages can be 
plotted in a modified AKF diagram (Fig. 
64) without violating the requirements of 
the mineralogical phase rule, suggesting 
that both the oxidative and nonoxidative 
mineral assemblages, viewed separately, 
closely approximate equilibrium assem- 
blages. The rarity of phyllosilicates 
crystallized directly from solution during 
oxidative diagenesis precludes at this 
time a decision as to whether the goe- 
thite-smectite or celadonite-chlorite tie- 
line is stable under those conditions, but 
the Fe-pigmented, oxidized smectites 
suggest that it is goethite-smectite. 

Bass, M. N., R. Moberly, J. M. Rhodes, 
C. S. Shih, and S. E. Church, Volcanic 
rocks cored in the central Pacific, Leg 
17, Deep Sea Drilling Project, Initial 
Reports, Deep Sea Drilling Project, 
17, 429-503, 1973a. 

Bass, M. N., R. Moberly, J. M. Rhodes, 
C. S. Shih, and S. E. Church, Volcanic 
rocks cored in the central Pacific, Leg 
17, Deep Sea Drilling Project, Trans. 
Am. Geophys. Union, 54, 991-995, 

Immiscibility and Ancient and 
Modern Volcanism 

C. Brooks and L. Gelinas* 

Archean volcanics from the Canadian 
Shield have long been the subject of 
major element, trace element, and iso- 
topic studies in this laboratory and the 
results have been reported in previous 
Reports. As part of this broad geochemi- 
cal program, we have often digressed 
into topics that show promise of furnish- 
ing information on the type and nature 
of ancient magmatism (e.g., the trace 
element and isotopic composition of relic 
pyroxenes from Archean volcanics (Year 
Book 73, p. 967). This study is another 
of these digressions, and in this case we 
are attempting to understand ancient 
magmatism in the light of immiscibility, 
a process of igneous differentiation which 
may leave distinct and unusual imprints 
in the products of volcanism. 

Many of the globule-like textures 
found in igneous rocks have been found 
to be demonstrably immiscible with their 
enclosing fractions (e.g., Philpotts and 
Hodgson, 1968; Ferguson and Currie, 
1972). Most of these examples come 
from mafic alkaline intrusives; however, 
surprisingly, some have been unearthed 
in tholeiitic associations. For instance 
Roedder and Weiblen (1971) report the 
occurrence of two immiscible silicate 
liquids (one enclosed as droplets in an- 
other and quenched to form two im- 
miscible glasses) in volcanic rocks from 
Hawaii, California, Greenland, and Can- 
ada. Similar observations have been 
made in the high-alumina olivine-tholei- 
ite of Hat Creek, California (Anderson 
and Gottfried, 1971) and in the tholei- 

* Genie Mineral, Ecole Poly technique, Mon- 
treal, Quebec, Canada. 



itic Deccan Traps (De, 1974). De (1974) 
also suggested that the trend of differ- 
entiation in the Upper Zone of the Lay- 
ered Series of the Skaergaard Intrusion 
was in part controlled by liquid immisci- 
bility, and this was subsequently verified 
experimentally by McBirney (1975). 
Further evidence linking immiscibility 
and tholeiitic magmatism has been ad- 
vanced by Gelinas et al. (1975), who 
explain the origin of variolitic lavas in 
tholeiitic suites of Canadian Archean 
volcanic piles by liquid immiscibility. 

It would seem, therefore, that some 
form of association exists between liquid 
immiscibility and tholeiitic magmas. In 
many cases this association links im- 
miscible splitting to only the late stages 
in magmatic history; however, the tho- 
leiitic Archean variolites of Canada, the 
komatiitic Archean variolites of South 
Africa (Ferguson and Currie, 1972) , and 
the Skaergaard intrusion of Greenland 
all involve immiscibility on a much 
grander scale. Are these examples merely 
artifacts of some peculiar geologic con- 
ditions, or is immiscibility a ubiquitous 
phenomenon that continually shadows 
tholeiitic magmatism, awaiting only the 
right physicochemical conditions before 
manifesting itself? 

The purpose of this study is to exam- 
ine this question by an appraisal of cer- 
tain ancient and modern volcanic suites 
representative of the main classes of 
volcanism (tholeiitic, calcalkaline, alka- 
line) with the aim of (1) evaluating the 
possible link between tholeiitic magma- 
tism and immiscibility, and (2) compar- 
ing the main classes of volcanism in the 
light of susceptibility to immiscibility. 
It will be revealed that not only is there 
a good case for tholeiitic magmas to be 
more commonly associated with immisci- 
bility than either the calcalkaline or al- 
kaline volcanics, but also that there is a 
stronger likelihood of encountering evi- 
dences of immiscibility in the more an- 
cient volcanics, as a result of a polarized, 
time-dependent proneness of certain 
magmas to undergo liquid segregation. 

Proposed Miscibility Gaps 

Grieg (1927) showed that of the major 
rock-forming oxides, Na 2 0, K 2 0, and 
A1 2 3 are miscible with Si0 2 whereas 
CaO, FeO, MgO, and Fe 2 3 are immisci- 
ble with Si0 2 at high Si0 2 concentra- 
tions. Grieg presented a pseudo-ternary 
diagram of (CaO+MgO+2FeO+ 
(TiO,)) versus (Na 2 0+K 2 0+Al 2 3 ) 
versus Si0 2 to show the limited extent 
of the primitive immiscibility field in 
the polycomponent system. In this sys- 
tem, Grieg's primitive field is restricted 
to a narrow zone high on the silica 
portion of the (CaO+MgO+2FeO+ 
(Ti0 2 ))-Si0 2 edge (Fig. 65). Although 
experimentally verified, Grieg considered 
immiscibility unlikely as a process in the 
origin of igneous rocks, on the bases 
that: (1) The experimentally-established 
minimum temperatures were too high 
(nearly 1700°C) ; (2) No igneous rocks 
approached the compositions found to 
show immiscibility in the laboratory; 
and (3) The textural evidence for im- 
miscibility in natural rocks could be 
readily assigned to other processes. It 
was not until Roedder (1951) described 
silicate immiscibility in the system leu- 
cite-fayalite-silica, and in the parent 
quaternary system K 2 0-FeO-Al 2 3 -Si0 2 
(1953), at much lower temperatures 
(minimum < 1100°), and in composi- 
tions much closer to igneous rocks, that 
the geological possibility of immiscibility 
became reasonable. 

In the development of knowledge con- 
cerning this second field of immiscibility 
(denoted PMG in Fig. 65), both Holgate 
(1954) and Roedder (1956) noted that 
two systems relevant to silicate melts 
display flat liquidus surfaces which ac- 
cording to Grieg (1927) are indicative 
of possible metastable fields of liquid im- 
miscibility. These systems are NaAlSi0 4 - 
FeO-Si0 2 (Bowen and Schairer, 1938), 
where the fayalite field displays a flat 
liquidus surface, and KAlSi 2 6 -CaMg- 
Si 2 6 -Si0 2 (Schairer and Bowen, 1938), 
where the plateau-shaped diopside field 
shows a peculiar inverse curvature. With 





CaO + MgO + EFeO+TiO 

Na,0 + K,0+AUO 


Fig. 65. Pseudo-ternary diagram showing location of the primitive and proposed miscibility 
gaps (latter denoted as PMG). Tie-lines within the PMG join experimentally established 
liquid-immiscible fractions (data from Roedder and Weiblen (R), 1970; Massion and Van 
Groos (M), 1973; McBirney (B), 1975). The open fields are for lunar glass inclusions 
(N = 50) of Roedder and Weiblen. 

the aid of Grieg's (1927) pseudo-ternary 
diagram, Holgate (1954) presented Roed- 
der's (1951) stability field of immisci- 
bility within the system leucite-fayalite- 
silica, together with the locations of the 
axes of the metastable miscibility gaps 
for the systems NaAlSi0 4 -FeO-Si0 2 and 
KAlSiO«-CaMgSi 2 6 -Si0 2 . The coinci- 
dence of these axes with Roedder's field 
of immiscibility was striking. 

The final validity of this field was 
obtained in the laboratory. Roedder and 
Weiblen (1970), Ferguson and Currie 
(1972), Massion and Van Groos (1973) 
and McBirney (1975) have obtained 
during experiments immiscible liquid 
pairs which plot within the PMG shown 
in Fig. 65. Hence, the PMG can be con- 
sidered confidently as a verified, geologi- 

cally-applicable two-liquid field. Pre- 
sumably, magmas with compositions 
plotting in or on the flank of the PMG 
would be prone to split into two liquids, 
and we use this property throughout our 
evaluation of the volcanic rock series. 

Because we use the PMG in our ap- 
praisal of volcanic rock suites, it is per- 
tinent here to consider some of the more 
important properties of this field as re- 
vealed on the pseudo-ternary diagram 
of Fig. 65. The PMG is an extended 
ellipsoid that roughly parallels the silica— 
cafemic oxides boundary. This means 
that a magma composition plotting on 
the alkaluminous side of the PMG would 
be driven into the PMG by any processes 
leading to a relative decrease in the al- 
kali and alumina contents of that 



magma. If this process also leads to a 
lower alumina-to-alkalies ratio, then im- 
miscible splitting is even further favored, 
although as pointed out by Watson and 
Dickey (1975) the ratio, more appropri- 
ately, should be alumina-to-alkalies plus 
water. These features have importance 
in the discussion to follow. 

Volcanic Rock Suites and the PMG 

Phenocryst-free, rapidly cooled vol- 
canic rocks can be used to tell much 
about the composition of their parent 
magma. In gathering data from the lit- 
erature we have undoubtedly included 
numerous phenocryst-bearing volcanic 
rocks; however, this is of secondary im- 
portance since we are primarily inter- 
ested in liquid compositions, and these 
may be either original melt compositions 
or compositions of late, residual liquids. 
The procedure used has been to assem- 
ble data from different tectonic regimes 
and to establish composite fields for the 
main classes of volcanism. 

The composite fields for modern vol- 

canics are plotted in Fig. 66. These 
volcanics define a vast domain on the 
right flank of the PMG, with their re- 
spective positions being controlled by 
their relative alkali-plus-alumina con- 
tents. Successively from the alkali- 
alumina corner towards the PMG, the 
suites encountered are alkaline, calc- 
alkaline, low-K calcalkaline, and tholei- 
itic. The only modern suite to overlap 
the PMG is the tholeiite suite, and on 
dissection, it is found to do so only for 
the more silicic low-K tholeiites of island 
arcs. No other modern volcanic suite 
approaches the PMG, although the low- 
K and normal calcalkaline fields closely 
parallel the boundary of the PMG. 

The composite Archean tholeiite and 
low-K calcalkaline suites are shown in 
Fig. 67B. The ancient tholeiite suite is 
essentially identical with the modern 
tholeiite suite (Fig. 67A) and most espe- 
cially with the low-K tholeiite suite of 
island arcs. Both of these suites overlap 
the PMG at higher Si0 2 concentrations, 
with the ancient suite displaying a slight 
trend to even higher Si0 2 contents than 


+ EFeO+Ti0 2 

Na ? 0+K ? 
+ AI 2 3 


Fig. 66. Composite fields for modern volcanics in relationship to the PMG. The alkaline, 
calcalkaline, low-K calcalkaline, and tholeiitic suites are based on 119, 73, 168, and 284 
analyses, respectively (data sources too numerous to identify). 




-£FeO+Ti0 2 


Na 2 0+K 2 
+AI 2 3 

Fig. 67. Composite fields for modern (A) and ancient (B) tholeiite and low-K calcalkaline 
fields with respect to the PMG. The fields are based on 345 (A) and 122 (B) analyses (data 
source, Gelinas et ah, 1975) . 

those encountered in the modern ex- 

The composite, ancient low-K calc- 
alkaline field differs from the modern 
field. While the modern field never ap- 
proaches the PMG, the ancient is posi- 
tioned significantly closer to the PMG 
and actually touches it at high Si0 2 
concentrations. For the most part the 
ancient volcanics possess lower alumina 
plus alkalies than their modern chemical 


It remains to relate these observations 
to immiscibility. On examining the abun- 
dance of textural evidence indicative of 
immiscibility, it is clear that Archean 
terranes contain many examples, whereas 
there are few in modern terranes. The 
best examples of immiscibility found in 
Archean terranes are the variolitic lavas 

of Canada and South Africa, and they 
are widespread. Is this a real phenome- 
non indicating a varying susceptibility 
of tholeiitic and ancient calcalkaline 
magmas to undergo liquid immiscibility 
with time? There is evidence to suggest 
that this may be so. 

Lower alumina and alkali contents re- 
sult in rock fields plotting in the PMG. 
It has been suggested (e.g., Hart et al., 
1970) that because of a higher geother- 
mal gradient, high rates of convective up- 
welling and reduced lithospheric thick- 
ness were probable in the Precambrian. 
Extensive and relatively shallow partial 
melting is a direct prediction of such 
thermal considerations, and this has al- 
ready been used in discussing the origin 
of komatiite (e.g., Viljoen and Viljoen, 
1969; Brooks and Hart, 1974; Cawthorn 
and Strong, 1974). Because extensive 
partial melting produces magmas with 



lower alkali and alumina contents, this 
model implies that the production of im- 
miscibility-prone magmas was more com- 
mon in the Archean than today. 

Further support for encountering im- 
miscibility-prone magmas early in 
Earth's history is provided by models of 
mantle water content over time. Strong 
and Stevens (1974), in discussing the 
change in Earth's behavior at the Ar- 
chean-Proterozoic boundary, postulate 
that the differences between Archean and 
younger sections result from a water- 
pressure effect on the peridotite solidus. 
With degassing of the mantle, smaller 
and smaller degrees of partial melting of 
mantle material occur. This is in direct 
accord with the previously discussed 
thermal model. However, water not only 
has a role in partial melting considera- 
tions, it has been shown to have a direct 
effect on the field of immiscibility. 

Holgate (1954) proposed that the field 
of immiscibility was enlarged at high 
water pressure, and as pointed out by 
Watson and Dickey (1975) water as a 
liquidus depressant has been found neces- 
sary in many cases of experimental veri- 
fication of the process (e.g., Philpotts 
and Hodgson, 1968; Philpotts, 1971; 
Ferguson and Currie, 1972). Arguments 
can be advanced for both wet and dry 
Archean volcanism; because (a) the 
Canadian Archean of 2.7 b.y. contains 
abundant andesites, and (b) the Archean 
tholeiites most closely resemble low-K 
island arc tholeiites, we are inclined to 
believe that the volcanism during the 
formation of the Archean metavolcanic 
belts may have been wet. Whatever the 
water content, however, this model, in- 
volving decreasing water contents in the 
accessible mantle with time, again favors 
immiscibility in Archean volcanic rocks 
relative to those of younger terranes. 

It would seem, on the basis of this 
evaluation of ancient and modern vol- 
canic suites, that a direct link between 
magmatism and immiscibility exists for 
certain tholeiitic and ancient calcalkaline 
magma compositions. These compositions 

are mostly in the intermediate-to-felsic 
range, a fact reminiscent of Grieg's 
(1927) experimental observation that 
CaO, FeO, MgO and Fe 2 H show incom- 
plete miscibility with Si0 2 at high Si0 2 

An important question relates to 
whether the potential immiscibility in 
ancient and modern tholeiitic magmas 
(and in ancient low-K calcalkaline mag- 
mas) is restricted to the late-stage resi- 
dual liquids only, or whether it involves 
much of the magma. The previous dis- 
cussion concerning lowering of relative 
alumina plus alkalies contents, and the 
role of water in magmas, is primarily 
applicable to the direct production of an 
immiscible-prone melt. However, differ- 
entiation characteristically enhances a 
residual magma in liquidus depressants 
such as iron and alkalies. Very high iron 
contents are especially necessary for 
successful immiscible splitting experi- 
ments in dry systems (e.g., Roedder and 
Weiblen, 1970; McBirney, 1975). Hence, 
where a tectonic environment does not 
allow direct enhancement of liquidus de- 
pressants in a primary magma, and that 
magma composition is not favorable to 
liquid immiscible splitting, it may still 
reach a two-liquid stage in its residual 
phases. We cannot yet distinguish with 
certainty which of these was responsible 
for the Archean variolites in the Abitibi 
belt, although their high iron contents 
removes the necessity of having wet tho- 
leiitic magmas. Because, however, the 
Archean variolitic lava compositions sug- 
gest that immiscibility is restricted to 
the higher Si0 2 levels of tholeiitic magma 
evolution and because the tholeiite com- 
posite field overlaps the PMG only at 
higher Si0 2 levels, it is probable that 
some magmatic differentiation is re- 
quired before a tholeiitic (or ancient, 
low-K calcalkaline) magma will split 
into two liquid fractions. This in turn 
favors Fe as a more important liquidus 
depressant than water in tholeiite 

The evaluation presented here is not 



complete. We are only now beginning to 
understand the secondary modification 
that pervades Archean volcanics, and as 
more and more data become available we 
may be able to extend or at least under- 
line the conclusions here presented. At 
this stage, however, we can state that 
the recent abundance of liquid-immisci- 
bility observations associated with tho- 
leiitic magmatism is not in conflict with 
what is experimentally known about im- 
miscibility. In fact, the reverse is true, 
and immiscibility considerations lead to 
a direct prediction that textural evi- 
dences of the phenomenon should be 
common in at least the more diversified 
members of the tholeiitic suites. Further- 
more, there is evidence to indicate that 
such textural evidences will be more 
abundant in ancient rather than in mod- 
ern terranes. Similar conclusions seem to 
apply to the distribution of peridotitic 
komatiites and ultramafic tuffs in time. 
Perhaps earth's mantle did leave a 
unique imprint at the Archean stage of 
continent evolution, an imprint which we 
are only now beginning to decipher. 


Anderson, A. T., and D. Gottfried, Con- 
trasting behavior of P, Ti, and Nb in 
a differentiated high-alumina olivine- 
tholeiite and a calc-alkaline andesite 
suite, Geol Soc. Am. Bull, 82, 1929- 
1942, 1971. 

Bowen, N. L., and J. F. Schairer, Crys- 
tallization equilibrium in nepheline- 
albite-silica mixtures with fayalite, 
J. Geol, 46, 397-411, 1938. 

Brooks, C, and S. R. Hart, On the sig- 
nificance of komatiite, Geology, 2, 
107-110, 1974. 

Cawthorn, R. G., and D. F. Strong, The 
petrogenesis of komatiites and related 
rocks as evidence for a layered upper 
mantle, Earth Planet. Sci. Lett., 74, 
369-375, 1974. 

De, Aniruccha, Silicate liquid immisci- 
bility in the Deccan Traps and its 
petrogenic significance, Geol. Soc. Am. 
Bull, 85, 471-474, 1974. 

Ferguson, J., and K. L. Currie, Silicate 
immiscibility in the ancient basalts of 
the Barberton Mountain land, Trans- 
vaal, Nature Phys. Sci., 235, 86-89, 

Grieg, J. W., Immiscibility in silicate 
melts, Am. J. Sci., 13, No. 73, 1-44, and 
13, No. 74, 13&-154, 1927. 

Hart, S. R., C. Brooks, T. E. Krogh, and 
G. L. Davis, Ancient and modern vol- 
canic rocks: a trace element model, 
Earth Planet. Sci. Lett., 10, 17-28, 

Holgate, N., The role of liquid immisci- 
bility in igneous petrogenesis, J. Geol, 
62, 439-480, 1954. 

Massion, P. J., and A. F. Koster Van 
Groos, Liquid immiscibility in sili- 
cates, Nature, Phys. Sci., 245, 60-63, 

McBirney, A., Differentiation of the 
Skaergaard intrusion, Nature, 253, 
691-694, 1975. 

Philpotts, A. R., Immiscibility between 
feldspathic and gabbroic magmas, 
Nature, Phys. Sci., 229, 107-109, 1971. 

Philpotts, A. R., and C. J. Hodgson, Role 
of liquid immiscibility in alkaline 
rock genesis, Twenty-third Int. Geol 
Congr., Vol. 2, pp. 175-188, 1968. 

Roedder, E., Low temperature liquid 
immiscibility in the system K 2 0-FeO- 
Al 2 3 -Si0 2 , Am. Mineral, 36, 282- 
286, 1951. 

Roedder, E., Liquid immiscibility in the 
system K 2 0-FeO-Al 2 3 -Si0 2 (abstr.), 
Geol. Soc. Am. Bull, 64, 1466, 1953. 

Roedder, E., The role of liquid immisci- 
bility in igneous petrogenesis: A dis- 
cussion, J. Geol, 64, 84-88, 1956. 

Roedder, E., and P. W. Weiblen, Silicate 
liquid immiscibility in lunar magmas, 
evidenced by melt inclusions in lunar 
rocks, Science, 167, 641-644, 1970. 

Roedder, E., and P. W. Weiblen, Petrol- 
ogy of the silicate melt inclusions, 
Apollo 11 and Apollo 12 and terrestrial 
equivalents, Proc. Second Lunar Sci. 
Conf., Geochim. Cosmochim. Acta, 
Suppl. 2, Vol. 1, MIT Press, 507-528, 



Schairer, J. F., and N. L. Bowen, The 
system leucite-diopside-silica, Am. J. 
Sci, 35 A, 507-528, 1938. 

Strong, D. F., and R. K. Stevens, Possi- 
ble thermal explanation of contrasting 
Archean and proterozoic geological 
regimes, Nature, 249, 545-546, 1974. 

Viljoen, J. J., and R. P. Viljoen, Evi- 
dence for the existence of a mobile 
extrusive peridotitic magma from Ko- 
mati Formation of the Onverwacht 
Group, Geol. Soc. S. Africa, Spec. Pub. 
2, pp. 87-112, 1969. 

Watson, E. B., and J. S. Dickey, Jr., 
Silicate liquid immiscibility in alka- 
line mafic magmas: evidence in a 
White Mountain magma series dike, 
with a discussion of its origin, Geo- 
chim. Cosmochim. Acta, in press, 1975. 

Early Evolution of the Central 
Andean Volcanic Arc 

David E. James, Christopher Brooks, 
and Arturo Cuyubamba 


We have reported previously (Year 
Book 72 and Year Book 73) on our 
efforts to derive a model for the pedo- 
genesis of the rocks of the earliest An- 
dean volcanic arc. Our attention in these 

studies has focused on two aspects of the 
initial phase of development of the 
Andean arc — the radiometric age and the 
Rb-Sr-K trace element geochemistry of 
the volcanic rocks. The radiometric age 
not only gives the time of inception of 
the Andean arc, but also provides infor- 
mation about plate motions in the Pa- 
cific during early Mesozoic time. The 
trace element geochemistry provides 
data on the depth to the Mesozoic sub- 
duction zone and on the chemical nature 
of the magmas of the primitive Andean 
arc relative to those of the more evolved 
arc of late Mesozoic and Cenozoic time. 
The earliest volcanic rocks of the 
Andean arc are preserved in a narrow 
belt situated on the coast of southern 
Peru (see Fig. 68 for outcrop map and 
sample localities). The rocks show con- 
siderable evidence of deposition near 
sea level as witnessed by pillows and 
intercalated shallow-water marine sedi- 
mentary rocks. The volcanic rocks are 
considered to belong to the Chocolate 
Formation described by Jenks (1948), 
although the rock of the Chocolate type 
locality near Arequipa is not volcanic 
(as Jenks erroneously thought) but is 
instead a volcanic sedimentary rock. The 
only Chocolate volcanic rocks known to 

17° 50' - 


i-'-m Chocolate volcanics 


70° 50' 

70° 40' 

Fig. 68. Outcrop map showing the distribution of early Jurassic volcanic rocks of the 
Chocolate formation. Sample localities and sample numbers are shown and sample groupings 
are given by Roman numerals. 



us occur along coastal Peru. They are 
primarily basalts and basic andesites 
(50%-57%SiO 2 ). 

A major difficulty in untangling the 
origin and history of the Chocolate vol- 
canics is the pervasive alteration that 
affects most of the samples. Rocks ap- 
pearing fresh in hand specimen are com- 
monly too altered to be of use geochemi- 
cally. As a rule, we have achieved con- 
sistent results by first screening samples 
on the basis of H 2 and C0 2 contents. 
We have confined our discussion to those 
samples with H 2 < 2.0% and C0 2 < 
0.2%. Our inability in previous years 
to obtain definitive results on the Choco- 
late volcanics stemmed principally from 
the paucity of samples that passed the 
screening criterion. For that reason we 
undertook during the past field season 
to obtain a substantially improved sam- 
ple collection. We have succeeded in 
tripling our useful data with the addi- 
tion of 13 samples that meet the H 2 0- 
C0 2 screening criterion. 

Analytical Results 

The analytical data discussed here are 
summarized in Table 10. Analytical pro- 
cedures have been described by us in 

previous Reports. All strontium isotopic 
measurements have been normalized to 
an assumed E and A standard value of 

Strontium isotopic composition. 87 Sr/ 
86 Sr ratios are plotted versus Rb/Sr in 
Fig. 69. Regression analyses have been 
made on several subgroupings of sam- 
ples and yield calculated ages ranging 
from 165 to 205 m.y. Our preferred age 
is 185 m.y., calculated for the group of 
samples with the least petrographic evi- 
dence of alteration. PE 120 appears fresh 
in thin section but is excluded because 
of its exceptionally high Rb/Sr ratio 
(2.1) and clear indication that its 87 Sr/ 
86 Sr ratio has begun to "droop" on the 
isochron. This is not an uncommon char- 
acteristic of rocks with high Rb/Sr. 

A histogram of initial 87 Sr/ 86 Sr ratios 
(excluding PE 120 and PE 316) is shown 
in Fig. 70. The mean ( 87 Sr/ 86 Sr) for an 
assumed age of 185 m.y. is 0.7038 and is 
identical to the average for normal is- 
land arcs. Initial ratios vary slightly 
with H 2 content, but the variation is 
not significant, suggesting that the ratios 
are little affected by alteration as meas- 
ured by H 2 0. 

Rb-Sr variations. Rb and Rb/Sr ratios 

TABLE 10. Analytical Data and Initial Sr-isotopic Ratios, Chocolate Volcanic Rocks of 
Southern Peru. Initial 87 Sr/ 8e Sr Ratios Calculated for T = 185 m.y. 

87 Sr/ 86 Sr 

87 Sr/ 86 Sr 

Sample No. 

Rb (ppm) 

Sr (ppm) 


(present day) 


H 2 0(%) 





0.70501 ± 10 







0.70519 ± 17 







0.70491 ± 11 







0.70411 ± 6 







0.71748 ± 8 







0.70751 ± 13 







0.70469 ± 5 







0.70620 ± 6 







0.70493 ± 5 







0.70585 ± 6 







0.70478 ± 7 







0.70457 ± 6 







0.70496 ± 8 







0.70556 ± 6 







0.70725 ± 7 







0.70488 ± 5 



* Not included in calculating average 87 Sr/ 86 Sr initial ratio. 




0.704 -; 


1 r 

J L 

J L 


Fig. 69. 87 Sr/ 86 Sr ratios of Chocolate volcanics plotted versus Rb/Sr. Solid circles indicate 
data used for regression analysis ; open circles and square are other data points. 

of the Chocolate volcanics are abnor- 
mally high for normal volcanic arc ba- 
saltic rocks. The high Rb is apparently 
due to the addition of alkalies during 
late-stage alteration (possibly by sea 
water) of the rocks at the time of their 
formation. That the alteration was 
roughly contemporaneous with extrusion 
is indicated by the fact that the radio- 
metric ages appear to be correct, being 
as great or greater than previous esti- 
mates based on stratigraphic correla- 
tions. In rocks with H 2 < 2.0%, only 
minor alteration due to weathering ap- 
pears to have occurred. 



N 3 



0.703 0.704 0.705 

(Sr 87 /Sr8 6 ) 

Fig. 70. Histogram of initial 87 Sr/ 88 Sr ratios 
of the Chocolate volcanics excluding PE 120 
and PE 316. Initial ratios based on 185 m.y. 



M - 








: '.. ".'■'■■' ' ' 

^V ' v 



Sr concentrations in the Chocolate 
volcanics are comparatively normal 
(~250-700 ppm) for volcanic arc ba- 
salts. The fact that the 87 Sr/ 86 Sr initial 
ratios are low suggests that the altera- 
tion does not greatly involve Sr which is 
relatively highly buffered against sea 
water (Hart et al., 1974) . 

A well-defined trend in the Rb-Sr data 
is obtained by plotting Rb/Sr vs. Rb as 
shown in Fig. 71. The good straight-line 
fit through the data could indicate that 
most of the change during alteration has 
been in Rb concentration, not in Sr. 
Only PE 161 and PE 316 lie signifi- 
cantly above the best-fit line through 
the data. Both of these samples have 
87 Sr/ 86 Sr ratios that plot significantly 
off the isochron of Fig. 69. 

Interpretation and Conclusions 

The earliest volcanic rocks of the cen- 
tral Andean arc were extruded about 185 
m.y. ago. They exhibit initial 87 Sr/ 86 Sr 
ratios that are identical to those meas- 
ured in rocks of oceanic island arcs. Low 
initial ratios have been preserved despite 
pervasive alteration that has added al- 
kalies (including Rb) to most of the sam- 
ples. Sr concentrations appear to have 
been little affected by alteration, al- 






Rb (ppm) 
Fig. 71. Rb/Sr vs. Rb. Straight-line fit excludes PE 313 and PE 161. 

though erroneous initial ratios may be 
more common in those samples with high 
Rb and other alkalies. 

The Chocolate volcanics and other 
Mesozoic volcanic sequences that fol- 
lowed (see Year Book 73) are notable 
among Andean rocks for their low 87 Sr/ 
86 Sr initial ratios. Late Cenozoic lavas 
of the same general region exhibit 87 Sr/ 
86 Sr in the range 0.705-0.708. We have 
shown previously that the high ratios of 
the late Cenozoic assemblages cannot be 
due to crustal contamination of a parent 
magma with 87 Sr/ 86 Sr near 0.704. Hence, 
the difference in isotopic ratios between 
the Mesozoic and Cenozoic sequences 
must reflect a genuine change in the 
petrogenesis of the respective magmas. 
We suggest that the earliest magmas of 
the Andean arc were derived from the 
descending plate and overlying oceanic- 
type asthenosphere or young continental 
lithosphere at depths less than 100 km. 
The later Cenozoic magmas, generated 
at depths of 150-175 km in the Benioff 
zone, probably derive from melting of 
ancient continental lithosphere immedi- 
ately overlying the descending slab and 
thus reflect the higher alkali abundance 
and aged isotopic character of that litho- 
spheric mantle. 


Hart, S. R., A. J. Erlank, and E. J. D. 
Kable, Seafloor basalt alteration: 
Some chemical and Sr isotope effects, 

Contrib. Mineral. Petrol., 44, 219, 
Jenks, W. F., Geology of the Arequipa 
quadrangle, Bol. Inst. Geol. Peru, 9, 
105, 1948. 

Strontium Isotopic Composition of 

Late Cenozoic Central Andean 

Volcanic Rocks: A Disequilibrium 

Melting Model 

D. E. James 


The late Cenozoic Arequipa and Bar- 
roso andesitic-dacitic volcanic rocks of 
southern Peru (geologic map and sam- 
ple localities shown in Fig. 72) form two 
geochemically distinct volcanic associa- 
tions that have been described elsewhere 
{Year Book 73, pp. 983-997, 1974). A 
puzzling aspect of the Rb-Sr isotope 
data for these two volcanic units is that 
they define two internally consistent and 
parallel pseudoisochrons that yield an 
age of 400 m.y. (Fig. 73) . The consistent 
variation of 87 Sr/ 86 Sr with Rb/Sr could 
be produced by at least three mecha- 
nisms: (a) crustal contamination, (b) 
small-scale inhomogeneity of the mantle 
source, and (c) disequilibrium melting 
of mantle material. The last two alter- 
natives are not necessarily exclusive. 

In the light of evidence presented pre- 
viously (James et al., 1974) crustal con- 
tamination is an unlikely cause of the 
observed strontium isotope variation. 




Fig. 72. Geologic sketch map showing principal late Cenozoic volcanic formations and 
sample localities: (1) Pliocene-Quaternary volcanic flows, mapped chiefly as belonging to the 
Barroso Group; (2) Pliocene-Quaternary rhyolitic ignimbrites of the Seneca formation; 
(3) Miocene-Pliocene rhyodactic ignimbrites of the Huaylillas formation; (4) non-volcanic 
formations. Sample localities are indicated by closed circles for the Barroso volcanics and open 
circles for the Arequipa volcanics. Very approximate regional distributions of the Arequipa 
and Barroso units are delimited by circles. Asterisks mark sites of andesitic strato-volcanoes. 

This conclusion is based chiefly on the 
large amounts of contaminant required 
and on the internal consistency of the 
isotopic variation within each of the two 
volcanic assemblages. Selective leach- 
ing of highly radiogenic phases from 
crystal rocks could greatly reduce the 
amount of contaminant required but 
could not plausibly account for the re- 
gionally independent internal consist- 
ency of the isotopic variation. 

James et at. (1974) interpreted the Sr 
isotopic variation within the Arequipa 
and Barroso volcanics as due to either 
small-scale inhomogeneities in the man- 
tle material at the source or disequilib- 
rium melting of comparatively homoge- 
neous source material. The latter process 
is usually described as melting of ho- 
mogeneous mantle without isotopic equi- 
libration between liquid and residual 
phases. Thus, as mineral phases with 










-_ 25 46 
23 ocR 

80 / 48 





Fig. 73. Variation of ^Sr/^Sr vs. Rb/Sr for Arequipa and Barroso volcanics. Open circles 
denote Arequipa volcanics; closed circles, Barroso volcanics. Lines define 400 m.y. pseudo- 
isochrons through each group of samples. 

greater or lesser abundances of Rb (and, 
hence, radiogenic Sr) enter the melt, the 
isotopic composition of the melt varies 
accordingly. A thorough discussion of 
disequilibrium melting models is given 
by Hofmann and Hart in this Report. 
James et al. (1974) hypothesized that 
the mantle source for the Arequipa and 
Barroso magmas is the continental litho- 
sphere immediately overlying the de- 
scending oceanic plate and situated at a 
depth of 150 to 175 km. By this hypothe- 
sis, the isotopic differences between the 
Arequipa and Barroso volcanics are the 
result of gross inhomogeneity in the con- 
tinental lithosphere at the site of magma 
generation. The isotopic discontinuity 
between the two volcanic units cannot 
be explained by normal disequilibrium 
melting. A disequilibrium melting model 

can be applied only to explain the inter- 
nal isotopic variation within each of the 
two volcanic assemblages. 

Concepts and Assumptions 

Numerous papers have been published 
recently on disequilibrium melting (see 
Hofmann and Hart in a companion re- 
port for a review of the literature and a 
complete list of references). The notion 
is both novel and attractive, albeit en- 
tirely unsubstantiated by laboratory ex- 
periments. The deficiencies of proposed 
disequilibrium melting models as applied 
to oceanic basalts are detailed by Hof- 
mann and Hart and appear highly for- 
midable. Nonetheless, if the Andean 
magmas are derived from solid conti- 
nental lithosphere, the total absence of 
incipient melt prior to actual magma 


formation should help significantly in 1969) . If this value is assumed and the 

maintaining long-term isotopic disequi- amount of mantle phlogopite increased 

librium between phases. Data by Hof- accordingly, the calculations summa- 

mann (this Report) support this asser- rized in this report will not be signifi- 

tion in that solid diffusion rates are cantly altered. 

found to be several orders of magnitude Within the framework of these as- 
lower than those of partially molten sumptions, I address the question of 
rocks. I am encouraged, therefore, to whether or not the observed variation 
apply the concepts of disequilibrium in strontium isotope ratios can be pro- 
melting to an explanation of the 400 ducecl by phlogopite-controlled disequi- 
m.y. pseudoisochrons exhibited by young librium melting of an appropriate man- 
Andean volcanic rocks, tie source. The objective is to compare 
The virtual absence of ground rules on calculated estimates of Sr isotope varia- 
disequilibrium melting creates an imme- tion based upon the degree of partial 
diate difficulty. It does seem clear, how- melting of phlogopite with observed 
ever, that advocacy of disequilibrium variations in the lavas exposed at the 
melting implies that at least certain surface, 
trace elements are held selectively within 

single phases and that those trace ele- Model LalcuLatwns 

ments enter the melt according to the Beswick (1975) lays the foundation 

degree to which their host phases par- for the calculations that follow and the 

ticipate in the melting process. Accord- reader is referred to his paper for de- 

ingly, and to limit an otherwise limitless tails. His principal arguments may be 

range of combinations and permuta- summarized as follows: 

tions, I make the following assumptions: 1. If K is held in the mantle as a 

1. All K and Rb in the mantle are stoichiometric component of phlogopite, 
contained in a single minor phase, phlo- the K concentration in a given melt will 
gopite, for depths of 100 to 200 km (see be controlled solely by the degree to 
Beswick, 1975). Radiogenic 87 Sr, pro- which phlogopite melts relative to other 
duced by decay of 87 Rb in the phlogo- phases involved in the production of 
pite, is assumed bound to the phlogopite melt. 

until the time of melting. By this as- 2. The K and Rb content of the melt 

sumption, excess 87 Sr in a melt is phlogo- can be equated with the mole fraction 

pite controlled. of phlogopite component in the melt if 

2. All additional Sr is contained in the distribution coefficient is known, 
clinopyroxene at mantle depths greater Beswick takes the distribution coeffi- 
than 100 km, where both plagioclase and cient to be 3. Given this, the liquid 
hornblende are unstable. fraction of phlogopite produced for a 

3. The composition of the mantle melt with a known concentration of K 
from which the magma is derived is that and known K/Rb ratio is given by 

of pyrolite or undepleted lherzolite (see v Rb /X Rh -J- *>Y Rb 3 

Ringwood, 1966; Shimizu, 1973; Shimizu j = r L + ^ A p ~ 6 

and Boyd, 1973) with K ~ 0.1% and —2 

K/Rb ~ 200-300. If all K is held in where X* b and X£ b are the mole frac- 

phlogopite, with 9.4% stoichiometric K, tions of Rb in p hlo r gopi te (P) an d liquid 

then phlogopite should constitute ~1% (L) , respectively. 

of the mantle. The actual K concentra- T h e total degree of me lting at the 

tions measured in phlogopite of ultra- S0U rce is obtained by multiplying the 

mafic inclusions found in kimberlites fraction of phlogopite liquid by that fac- 

and elsewhere average about 8% (Orif- tor necessary to achieve the appropriate 

fin and Murthy, 1969; Allsopp et al, dilution of K in the final melt. 



87 Sr/ 8Q Sr enrichment. Initial 87 Sr/ 86 Sr 
ratios of the Arequipa and Barroso vol- 
canics based upon a 400 m.y. pseudo- 
isochron are 0.7061 and 0.7044, respec- 
tively. Irrespective of the amount of 
disequilibrium melting and assuming all 
Rb is contained in phlogopite, the initial 
ratios must reflect the bulk Sr isotopic 
composition of the source region exclu- 
sive of phlogopite. Thus, in the discus- 
sion that follows, the 87 Sr/ 86 Sr ratios of 
all mineral phases except phlogopite are 
assumed equal to the initial ratio and all 
excess 87 Sr is assumed to come from 

Pertinent analytical data on the Are- 
quipa and Barroso volcanic rocks are 
summarized in Table 11, part A. The K 
contents and K/Rb ratios of these rocks 
(shown plotted in Fig. 74) are repre- 
sented as two end-point pairs for each 
of the two volcanic assemblages. The 
end-point values are used to calculate 
the fractions of phlogopite melt and to- 
tal melt given in Table 11 , part B. 

To calculate the amount of 87 Sr en- 
richment due to the melting of phlogo- 
pite, I assign a Rb-Sr age of 400 m.y. to 
the phlogopite and assume a present-day 
Rb/Sr ratio of 13.9 (Griffin and Murthy, 
1969). (The choice of a particular Rb/ 
Sr ratio does not affect the calculations 
because the 87 Sr/ 86 Sr ratio varies as a 
linear function of the Rb/Sr ratio. Thus 
phlogopite with greater or less Sr con- 
centrations will have proportionately 
lower or higher present-day 87 Sr/ 86 Sr 

1 1 


\ ° 
> o 




o\ O 

\ \ 

\ • o \ 

• \ \ 



1 1 




K (%) 

Fig. 74. Variation of K/Rb versus K for 
Arequipa and Barroso volcanics. Open circles 
denote Arequipa volcanics; closed circles, Bar- 
roso volcanics. Asterisks denote end-points 
used for calculations summarized in Table 11. 

ratios.) Taking 1.39 X 10- n /yr to be 
the decay rate of 87 Rb and assuming an 
initial 87 Sr/ 86 Sr ratio of 0.706, the 
present-day 87 Sr/ 86 Sr ratio in the phlo- 
gopite will be 0.925. 

Calculated values for 87 Sr/ 86 Sr based 
upon the assumptions already outlined 
are summarized in Table 11, part B. The 
ratios are obtained by assuming simple 
mixing between melted phlogopite and 
the remaining component of the melt 
and taking the bulk Sr concentration of 
the melt to be equal to the measured Sr 
concentrations given in Table 11, part A. 

The first column of 87 Sr/ 86 Sr values is 
calculated assuming that all phlogopite- 

TABLE 11 A. Summary of K-Rb-Sr Analytical Data for Late Cenozoic Volcanic 

Rocks of the Central Andes* 


Rb (ppm) 


Sr (ppm) 

87 Sr/ 8e Sr 

( 87 Sr/ 8e Sr) initial 

Arequipa Volcanics 














Barroso Volcanics 














* a, b, c, and d are the end-members of the Arequipa and Barroso volcanics. 



TABLE 11B. Compilation of Disequilibrium Melting Model Calculations for Late Cenozoic 

Volcanic Rocks of the Central Andes* 

Total melt (%) Phlogopite melt (%) 

7 Sr/ 8e Srf 

37 Sr/ 86 Sr$ 

Sr cP x (ppm) 










Source I: Mantle = 1% phlogopite with K/Rb = 300 

a. 4.5 0.77 

b. 3.3 0.92 

c. 5.5 0.94 

cannot be obtained from source with K/Rb >255 

Source II: Mantle = 1% phlogopite with K/Rb = 250 

a. 3.7 0.62 

b. 2.9 0.80 

c. 4.9 0.83 

d. 3.6 0.99 

Source III. Mantle = 1% phlogopite with K/Rb = 200 

a. 2.4 0.40 

b. 2.3 0.63 

c. 3.9 0.66 

d. 3.1 0.86 

* Assuming K phlog . = 9.4%; ( 87 Sr/ 86 Sr) ph , g. = 0.925; (Rb/Sr) P h.o g . = 13.9; and Sr ph ,o g . = 18 
ppm (Griffin and Murthy, 1969). a, b, c, and d are the end-members of the Arequipa and 
Barroso volcanics. 

t Ratios calculated assuming all phlogopite-held Sr enters the melt. 

% Ratios calculated assuming phlogopite-held Sr enters the melt in proportion to the fraction 
of phlogopite melted. 

























held Sr enters the melt. This is roughly 
equivalent to assuming an extremely 
low Sr distribution coefficient between 
crystal and melt for phlogopite. For the 
Arequipa volcanics in this instance, the 
best agreement between observed and 
calculated 87 Sr/ 86 Sr ratios is obtained 
for source I (K/Rb = 300). For the 
Barroso volcanics, the best agreement is 
obtained for source II (K/Rb = 250). 

The second column of 87 Sr/ 86 Sr values 
is calculated assuming that phlogopite- 
held Sr enters the melt in proportion to 
the amount of phlogopite melted. In this 
case, the calculated 87 Sr/ 86 Sr does not 
vary with K/Rb ratio and thus provides 
no constraint on the K/Rb of the source. 

In both sets of calculations the agree- 
ment with observed 87 Sr/ 86 Sr ratios is 
good, but the agreement is somewhat 
better for that case when all phlogopite- 
held Sr enters the melt. In all instances 
there appears to be a rather small degree 
of partial melting, typically 3.5% to 
5%. This may be near the lower limit 

of melt fraction necessary for magma 

The calculated 87 Sr/ 86 Sr ratios cannot 
be made to correspond more closely to 
the observed values by changing the as- 
sumption of a 400 m.y. age for the 
phlogopite. An assumed age of more 
than 400 m.y. will result in calculated 
87 Sr/ 86 Sr ratios being too high; a lower 
age will further compress the already 
too small interval between calculated 
end-member isotopic ratios. 

Sr variation. Both the Arequipa and 
Barroso volcanics exhibit high Sr con- 
centrations, much higher than can be 
obtained by melting of pyrolite or un- 
depleted lherzolite. The latter typically 
contains 10-15% clinopyroxene with 100 
to 120 ppm of Sr (Shimizu, 1973). It 
seems necessary to postulate, therefore, 
that the Sr concentrations in the clino- 
pyroxene of the source region are higher 
than those measured for clinopyroxene 
from lherzolite. 

The requisite Sr concentrations of the 



clinopyroxene can be calculated by the 
amount of clinopyroxene in the mantle, 
the fraction of total melt, and the dis- 
tribution coefficient for Sr between solid 
and melt. For simplicity it is assumed 
that phases melt in proportion to the 
proportion in which they are present. 
Given 12% clinopyroxene in the mantle 
and a distribution coefficient, K = 0.067, 
the strontium concentration of clino- 
pyroxene is given by : 

Ccpx = Cmelt (F + (x - f)K<**)/x 

where C cpx and C me it are the concentra- 
tions in the clinopyroxene and melt, re- 
spectively; F is the fraction of melt; x 
is the fraction of clinopyroxene origi- 
nally present; / is the fraction of clino- 
pyroxene entering the melt, and K cpx is 
the distribution coefficient. 

The results of the calculations sum- 
marized in Table 11, part B, show the 
Sr concentrations to be about twice as 
high as those normally measured in 
clinopyroxenes of undepleted lherzolites. 
This conclusion is not significantly al- 
tered by assuming larger relative frac- 
tions of clinopyroxene entering the melt. 

Summary of Conclusions 

The pseudoisochrons observed in 
young andesitic-dacitic volcanic rocks 
of the central Andean arc can be ex- 
plained by postulating disequilibrium 
melting of a 400 m.y. mantle source con- 
taining about 1% phlogopite. This con- 
clusion, if correct, implies that excess 
radiogenic 87 Sr produced in phlogopite 
has not migrated from the lattice over 
the past 400 m.y. period. This seems 
possible only if the rock in which the 
phlogopite occurs has remained solid 
during that period of time (Hofmann 
and Hart, this Report). This require- 
ment is consistent with the model pre- 
sented by James et al. (1974) in which 
the Arequipa and Barroso magmas were 
derived from ancient continental litho- 


Allsopp, H. L., L. O. Nicolaysen, and P. 
Hahn-Weinheimer, Rb/K ratios and 
Sr-isotopic compositions of minerals 
in eclogitic and peridotitic rocks, 
Earth Planet. Sci. Lett, 5, 231, 1969. 

Beswick, A. E., K and Rb relations in 
basalt and other mantle derived ma- 
terials: Is phlogopite the key?, sub- 
mitted to Geochim. Cosmochim. Acta., 

Griffin, W. L., and V. Rama Murthy, 
Distribution of K, Rb, Sr, and Ba in 
some minerals relevant to basalt gen- 
esis, Geochim. Cosmochim. Acta, 33, 
1389, 1969. 

James, D. E., C. Brooks, and A. Cuyu- 
bamba, Andean Cenozoic volcanism: 
Magma genesis in the light of Sr- 
isotopic composition and trace ele- 
ment geochemistry, in Carnegie Inst. 
Wash. Year Book 73, p. 983, 1974. 

Ringwood, A. E., The chemical composi- 
tion and origin of the earth, in Ad- 
vances in Earth Science, P. M. Hur- 
ley, ed., p. 287, MIT Press, Cam- 
bridge, 1966. 

Shimizu, N., Geochemistry of ultramafic 
inclusions from Salt Lake Crater, 
Hawaii, in Carnegie Inst. Wash. Year 
Book 72, p. 262, 1973. 

Shimizu, N., and F. R. Boyd, Trace ele- 
ment contents of clinopyroxenes from 
garnet lherozlites in kimberlites, in 
Carnegie Inst. Wash. Year Book 72, 
p. 272, 1973. 

Anomalous Island Arc Asthenosphere 
and Continental Growth 

/. Selwyn Sacks 


Lithospheric plate motions are influ- 
enced by the rigidities and viscosities of 
those parts of the earth involved in the 
tectonic process, especially the crust and 
upper mantle. Viscosity in this region is 
difficult to determine directly but may 
be estimated from the anelasticity (Q _1 ) 
by means of empirical relationships. 



Anelasticity studies have shown that 
there are large regional variations in the 
thickness of the high-viscosity litho- 
sphere. For example, the lithosphere be- 
neath South America is more than 300 
km thick, that beneath Japan only 70 
km thick. The study described here may 
explain how new lithosphere is generated 
during the subduction process. 

The anelasticity structure in island 
arc regions is shown in Fig. 75 (Sacks 
and Okada, 1974). The details of the 
structure in Fig. 75 are represented 
quantitatively by appropriate values of 
Q. It is known from observation that 
earthquakes occur only in high-Q re- 
gions. In the study by Sacks and Okada 
(1974), the shallower part of the as- 
thenosphere to a depth of about 250 km 

was modeled as laterally homogeneous 
with a Q p of 400 (see Fig. 75). It was 
observed, however, that seismic rays 
which spent significant time near the 
apex of the wedge formed by the hori- 
zontal and subducting lithospheres {A 
in Fig. 75), had a higher apparent Q 
than predicted by this model. The study 
reported here was undertaken to deter- 
mine the Q structure of this anomalous 
region in some detail, and to investigate 
the effect of this structure on the stress 
state of the subducting lithosphere. A 
model is presented for the evolution of 
continental lithosphere based upon the 
revised Q structure. 

Data. The data analyzed in this study 
came from the large dynamic range, wide 
frequency band tape recording seismo- 


2 00 







1400 KM 


. 2-2.2-2. 5' .' • 2(;3l- 1 .' uuu l 3£3423f 31 f ." 

21 1-'-53732f55* . *. ' • 


— 100 




— 500 


Fig. 75. Q p beneath Japan determined by Sacks and Okada (1974). Dots indicate regions of 
high Q (1000-3000); slanted lines, intermediate Q (200-500); and hatching, low Q (50-100). 
It was observed that ray paths through region A had higher Q than predicted. 



graphs (Sacks, 1966) operating near the 
Pacific coast of Hokkaido (KMU) and 
in central Honshu (MAT), Japan. The 
KMU instrument is located near the 
trench, and the MAT seismograph is 
near the volcanic front (see Fig. 79). 
The wide frequency range of the seismo- 
graphs used (1/20 to 10 Hz) facilitated 
accurate determination of the spectral 
stops of the seismic arrivals. Spectral 
ratio techniques used here are independ- 
ent of the earthquake's radiated spec- 
trum (Sacks, 1968). 

Q in the anomalous region. In general, 
seismic rays from an earthquake to a 
seismograph will traverse not only the 
region of interest (the near-trench sec- 
tion of the asthenosphere) but also other 
regions which will affect the spectrum of 
the seismic arrival. The Q values of a 
particular region are calculated from the 
overall, or apparent, Q value of the total 
path using the following identity : 

T/Q = Z U/qt 

i = 1 

where T is the total seismic travel time; 
Q is the apparent Q value for the total 
path; U is the travel time through the 
ith. section of the path, e.g., through the 
asthenosphere; g* is the apparent Q 
value of the ith. section; and n is the 
number of sections considered. 

Since all seismic waves have to tra- 
verse the lithosphere beneath Japan, the 
Q value and thickness of this region first 
had to be determined. The technique 
used is illustrated in Fig. 76. Seismo- 
grams recorded at MAT from earth- 
quakes occurring in the subducting plate 
at depths ranging from 18 to 300 km 
were analyzed. The apparent Q values 
for these various paths were determined. 
For ray paths entirely within the (con- 
tinental) lithosphere, Q values of about 
1000 were obtained, nearly independent 
of the depth of the earthquake. How- 
ever, for earthquakes at greater depths 
with ray paths lying partly in the as- 
thenosphere, the apparent Q was signifi- 
cantly lower. Because the velocities in 
the asthenosphere are slightly lower 

Fig. 76. Schematic of subduction zone showing ray paths studied to determine Q of the 
lithosphere. Crosses indicate earthquakes. There is a significant difference in the apparent Q 
between the two lower paths. 



than those in the overlying lithosphere, 
ray paths tend to plunge fairly deeply 
into the asthenosphere once they enter 
it. Thus a substantial part of the path is 
in the lower Q medium even if the earth- 
quake depth is only slightly greater than 
the thickness of the continental litho- 
sphere (see Fig. 76). It was found that 
for earthquake depths less than 60 km 
the Q values were lithospheric, i.e., about 
1000; but at depths greater than 77 km 
the apparent Q of the path had dropped 
significantly. The results from two 
earthquakes straddling the base of the 
lithosphere are given in Table 12. It can 
be seen that the earthquakes have simi- 
lar epicenters (only the depth being dif- 

ferent), and are of similar magnitude so 
that the same suite of seismograph sta- 
tions would be used for the hypocenter 
location, assuring that their relative lo- 
cation is likely to be accurate. Based on 
results from these two earthquakes, the 
thickness of continental lithosphere was 
found to be 70 km, and it was found to 
have a Q value of 1000. 

The Q structure of the asthenosphere 
was determined by studying the ray 
paths shown in Fig. 77. The apparent Q 
measurements for the paths shown in 
Fig. 77 are given in Table 13. Figure 78 
illustrates the anelasticity structure 
found for the upper 280 km of the as- 
thenosphere. The wedge-shaped section 

TABLE 12. Hypocentral Data, Travel Times, and Anelasticity Results for Earthquakes Just 
Above and Below the Base of the Continental Lithosphere 




P time 

S time 


Q P apt 

25 May 1969 
25 April 1971 










Fig. 77. Ray paths studied to determine Q structure of the asthenosphere. Crosses indicate 
earthquakes. The two paths from the shallower earthquake in the subducting plate gave higher 
Q to KMU than to MAT (after correcting for the effects of continental lithosphere). The 
deep event gave the same Q value to MAT as the shallower event did. 



(Km) 120 


Fig. 78. Q model for the Japan region. The hatched region has an anomalously high Q for 
asthenosphere, Q p = 500. Normal oceanic asthenosphere has Q p = 200. In an earlier study, 
before the existence of the anomalous region was recognized, Sacks and Okada (1974) sug- 
gested a value of 400 for a homogeneous asthenosphere. 

of asthenosphere bounded above by the 
continental lithosphere and below by the 
subducting lithosphere to a depth of 
about 110-120 km has a Q p of about 
500. The asthenosphere to the west of 
this wedge, to a depth of 300 km (the 
limit of this study), has a Q p value of 
200. The Q = 200 region appears to be 
fairly homogeneous, both laterally and 
vertically. The ray paths used for deter- 
mining Q of the asthenosphere are shown 
in Fig. 76. Table 13 shows the Q calcu- 
lated for the asthenosphere at various 
depths. Even for the shallowest earth- 
quake still below the continental litho- 
sphere, 77 km, a Q p value of 190 was 
found. This value is less reliable than 
those from the deeper earthquakes, since 
the proportion of the total travel time 
spent in the asthenosphere is sensitive to 
the velocity structure, which is not well 

It is important to know which of these 
two regions in the asthenosphere beneath 
Japan is most similar to the astheno- 
sphere under the ocean. In an earlier 
study, Sacks and Okada (1974) com- 
pared the spectra observed at KMU and 
MAT from a distant deep earthquake 
(Fiji). The Q of the asthenospheric 
wedge beneath MAT is similar to that 
in the same depth range in the astheno- 
sphere beneath the Japan trench, i.e., the 
Q p = 200 region appears to be a typical 
sub-oceanic asthenosphere. 

Seismicity. Higher Q values are com- 
monly assumed to indicate higher vis- 
cosities and the presence of a high-Q 
zone sandwiched between the continen- 
tal lithosphere and that of the descend- 
ing plate should be detectable using 
other methods sensitive to the state of 
stress in the region. One obvious indica- 
tor of the state of stress is the seismicity. 

TABLE 13. Asthenosphere Anelasticity Results for Paths Shown 

in Fig. 77 



Q p Apparent 

Q Asthenosphere 
















Figure 79 shows a map of the Japan re- 
gion with the plans of the seismicity 
sections studied superimposed. Hypo- 
centers determined from the Japan 
Meteorological Agency (JMA) network 
of seismographs (about 120 sites) were 
plotted for each section. Hypocenters 
determined from the JMA net are sys- 
tematically displaced from those deter- 
mined using teleseismic stations (Utsu, 
1971). This effect presumably is related 
to high seismic velocities in the down- 
going plate. Nonetheless, because of the 
high density of local seismic stations, 
the relative hypocenter locations deter- 
mined by the JMA network are reliable 
even if absolute locations are not. Figure 

80 shows the earthquakes in one of the 
sections (10) under the continental re- 
gion. It shows the normal pattern of 
island arc seismicity; i.e., in the subduc- 
tion zone, some seismicity in the conti- 
nental lithosphere, and an earthquake- 
free zone between these two regions. 
Figure 81 is a plot of the number of 
earthquakes occurring in the subduction 
zone as a function of depth of some con- 
tinental sections. Most of the earth- 
quakes occur in the zone of interaction 
between the subducting and continental 
lithospheres, i.e., in the upper 70 km or 
so, a value consistent with that deter- 
mined by Q. The seismicity in the sub- 
ducting plate below the continental 


40° + 

^50 c 


AO c 




Fig. 79. Map of Japan showing plans of seismicity sections. The sections are normal to the 
ocean trenches. Sections 1, 2, 13, 14, 15, and 16 are considered oceanic; and sections 3, 4, 4.5, 
5, 6, 7, 8, 9, 10, 11, and 12 are considered continental. The seismograph stations KMU and 
MAT are indicated. % 



lithosphere is generally much reduced, downgoing plate in the depth range 70- 

In this region the subducting plate is in 120 km. In this depth interval, the 

contact with relatively low- viscosity as- downgoing plate is adjacent to the 

thenosphere. There appears to be a sec- anomalously high Q region of the asthen- 

ondary region of high seismicity in the osphere. Seismicity sections outside the 









300 200 100 

I 1 !— 

1 1 21 112 4 1 


11 11 
112 11 
1 1 

1 212 1 1234B57C62 

I 1 1 1 13 1 

II 11 1 12 13464686683331_ 

2 11256F**D69* 12532 
1 272385033 1 3 
1233778564521 2 1 

1 1 1413 5 1113 
1 11122621 
231 3131 4 21 
1111 2 2 

1 I 






Fig. 80. Seismicity section 10, showing the dipping seismic plane, aseismic wedge, and 
earthquakes in the horizontal continental lithosphere. Only earthquakes in the upper 200 km 
are shown. 

















W-A \ 

No. 10 

No. 8 














Fig. 81. Number of earthquakes in the subduction zone vs. depth for sections 8 and 10 
(Fig. 78). The highest seismicity occurs in the upper 70 km where there is a lithosphere- 
lith osphere interaction, but there is a secondary peak in the 80-120 km depth range. 



continental area, i.e., north of Hokkaido 
or in the Izu-Bonin arc, do not show this 
secondary zone of seismicity. Figure 82 
compares the seismicity (in the subduc- 
tion zone) beneath continental Japan 
with that of adjacent regions. The earth- 
quake occurrence in these two regions is 
generally similar except in the depth 
range 80-120 km, in which there is sig- 
nificantly higher activity in the conti- 
nental group. 

Pressure axes. A further manifestation 
of the state of stress is the direction of 
the pressure axes of the force system 
causing the earthquakes. The pressure 
axes can be determined from observa- 
tions of first-motion and shear-wave 
polarization from both local and tele- 
seismic seismograms. Ichikawa (1971) 
studied the focal mechanisms of hun- 
dreds of earthquakes occurring in the 
Japan arcs. The nodal planes were 
picked by computer, which also assessed 

the goodness of fit. Ichikawa's summary 
of the results is shown in Fig. 83. At 
depths where there is essentially a litho- 
sphere-lithosphere interaction (less than 
70 km), pressure axes are either hori- 
zontal or normal to the dipping plane. 
This holds true for all sections of the 
Japan arcs. It is found also that at 
depths greater than 120 km the pressure 
axes generally are in the direction of the 
dipping plane. However, beneath the 
continental areas, the pressure axis re- 
mains approximately horizontal and 
does not switch to the in-plate type of 
the deeper earthquake zone until a depth 
of about 120 km, which is well below the 
continental lithosphere. 


There are indications from three inde- 
pendent techniques that there is a zone 
of asthenosphere with higher than nor- 


1 1 1 1 




£ 100 



w 50 







£ io 




Continent -— 3,6,8,12 


Ocean 1,2,13,14 

1 1 1 1 

1 1 



85 115 




Fig. 82. Comparison of the number of earthquakes in the subduction zone beneath the 
continental areas 3, 6, 8, 12 (Fig. 78) and beneath oceanic areas adjacent to the continent, 
1, 2, 13, 14 (Fig. 78). The seismicity is generally similar except for the depth interval, 100-130 
km, in which there are more earthquakes in the continental region. 



,.\xV" , : ^* - 



Fig. 83. Ichikawa's (1971) summary of the axes of maximum compression in the Japan arcs 
derived from focal mechanism studies. The subduction zone earthquakes at depths less than 
about 80 km all have pressure axes which are horizontal or normal to the dipping seismic 
plane. In the continental sections, C, D, E, F, G, this same pressure axis direction persists to 
a depth of about 120 km. Below this depth, the pressure axes lie in the seismic plane. 

mal ridigity below the continental litho- 
sphere near the trench. This zone ap- 
pears to exist beneath continental-type 
regions only (based on seismicity con- 
siderations), and may not occur in 
ocean-ocean type arcs such as Izu- 

The higher Q and higher rigidity of 
the anomalous region of asthenosphere 

may be due to a smaller degree of par- 
tial melting and hence lower tempera- 
tures than in normal asthenosphere. We 
conjecture that the cold subducting plate 
freezes out asthenosphere at the apex of 
the wedge (nearest the trench) where 
the circulation of asthenosphere material 
would be restricted. An evolutionary 
model for this process is shown in Fig. 



Motion of continent 
away from trench 

Fig. 84. Evolutionary model for a thick lithosphere under continents. The subducting plate 
is cold relative to the asthenosphere and freezes out a small segment, shown shaded (a). This 
segment will be eroded at that face (unbounded by lithosphere) by circulation of the adjacent 
(partially molten) asthenosphere. If there is any motion of the continental plate away from 
the trench, the higher rigidity wedge (shaded, b) will grow. 



84. In the absence of motion other than 
subduction, a small wedge will be frozen 
out of the asthenosphere but will not 
grow because the side away from the 
trench will be eroded by the circulating 
asthenosphere. If there is motion of the 
continental side away from the trench 
for any reason, this higher wedge will 
grow, resulting in a thickening of con- 
tinental lithosphere. Supporting evidence 
for this motion comes from Proverbio 
and Quesada (1974), who used astro- 
nomical observations from international 
latitude stations to determine that Ja- 
pan and Europe are approaching one 
another at a relative velocity of 1 cm/ 
year. It is possible that a mechanism 
such as this can be invoked to explain 
the thick lithospheres under the conti- 
nents (e.g., Sacks and Okada, 1974). In 
this model, subduction at continental 
borders will cause an increase in the 
volume of continental lithosphere. Over 
the span of earth history, the total vol- 
ume of lithosphere added to the conti- 
nents will likely be large. 


Ichikawa, M., Re-analysis of mecha- 
nisms of earthquakes which occurred 
in and near Japan and statistical 
studies on the nodal plane solutions 
obtained, 1926-1968, Geophys. Mag., 

Proverbio, E., and V. Quesada, Secular 
variations in latitudes and longitudes 
and continental drift, J. Geophys. 
Res., 79, 4941-4943, 1974. 

Sacks, I. S., 1966. A broad-band large 
dynamic range seismograph. In: J. S. 
Steinhart and T. J. Smith (Editors), 
The Earth Beneath the Continents. 
Geophys. Monogr. 10, A.G.U., Wash- 
ington, D.C., p. 543. 

Sacks, I. S., Q for P waves in the mantle, 
in Carnegie Inst. Wash. Year Book 66, 
pp. 28-31, 1968. 

Sacks, I. S., and H. Okada, A compari- 
son of the anelasticity structure be- 
neath western South America and 

Japan, Phys. Earth Planet. Interiors, 
Utsu, T., Seismological evidence for 
anomalous structure of island arcs 
with special reference to the Japanese 
region, Rev. Geophys. Space Phys., 9, 
839-890, 1971. 

Determination of the Subducting 

Lithosphere Boundary by Use of 

Converted Phases 

J. Arthur Snoke and I. Selwyn Sacks 


The most commonly used technique 
to determine the upper boundary of an 
underthrusting plate in subduction zones 
is to correlate that boundary with a dip- 
ping plane of seismicity. Such a seismic 
plane is often well defined in island arcs, 
such as Japan, but is not well defined in 
cordilleran arcs, such as western South 
America. In the past few years two inde- 
pendent techniques have been developed 
to define the dipping interfaces of de- 
scending slabs. These techniques give 
results in good agreement with those ob- 
tained using seismicity and can there- 
fore be used to define dipping interfaces 
in regions where the seismicity is either 
sparse or scattered. One technique em- 
ploys the ScSp phase (Okada, 1971, 1973, 
1974), and the other uses a class of 
anomalous high-frequency shear arrivals 
that has been described previously 
(Snoke et al, 1973, 1974a, 19746, 1974c; 
Isacks and Barazangi, 1973) . In this re- 
port the two techniques are described 
briefly, and comparison is made of both 
their assumptions and their potential as 
determinators for the location of the 
upper boundary of subducting plates. 

ScSp. An observed longitudinal-wave 
precursor to the ScS phase at Japanese 
and western South American stations 
has been identified as an ScSp phase; 
that is, a P phase converted from an ScS 
phase at a dipping interface (Okada, 
1971, 1973, 1974). All observations ap- 
pear to be consistent with this appar- 
ently unique interpretation (see Fig. 85) . 


,>«**» SOUTH A ^«<CA 

▼ Station 

• Hypocenter 
++ Earthquakes 


Fig. 85. A schematic model showing the paths of the ScSp phase (solid line) and the ScS 
phase (dashed line) . 

The phase is most clearly seen on short- shows a case where a seismic plane is not 
period recordings of large, deep-focus well denned. (See Okada, 1974, for fur- 
earthquakes, ther discussion.) 

The time interval between the ScS The part of the interface sampled is 
and ScSp arrivals from an event produces approximately on a line perpendicular 
a locus of possible ScS-to-P conversion to the trench and passing through the 
points. The observed time difference, station. This follows from the fact that 
AT = Tscs — Tscsp, is not sensitive to the ScS ray path is approximately along 
the details of the velocity contrast be- a radius vector of the earth {6 S ~ in 
tween the downgoing plate and the sur- Fig. 88), and the scattering plane de- 
rounding mantle. Within experimental fined by an incident ray along a radius 
error, AT is a constant for a given sta- vector and the normal to the interface is 
tion. Amplitude studies by Okada (1973, the plane which intersects the interface 
1975) indicate that the angle between along its maximum dip. To test the 
the incident ray and the normal to the validity of this approximation we ex- 
interface is near the critical angle. amine the ray path for the ScSp ob- 

Figures 86 and 87 show vertical seis- served at NNA, Peru, from an earth- 

micity cross sections taken perpendicu- quake on 5 March 1965. This event is 

lar to the trench and passing through atypical of those studied because of the 

two stations at which ScSp has been large epicentral distance (A = 19.6°) 

observed, KMU in Hokkaido, Japan, and because a 1} the angle between the 

and NNA in Peru, South America. Pro- back-azimuth and the horizontal projec- 

jections of surfaces on which AT = con- tion of the maximum-dip direction (see 

stant are drawn, and the projection cor- Fig. 88), is about 80° (Fig. 89). The 

responding to the observed AT is angle, a 2 , between the projected maxi- 

hatched. On each figure a circle shows mum-dip direction and the projection 

the preferred conversion region predicted of the outgoing P, is given by 

on the basis of amplitudes. In island • a 

i ■ i . I, i . v p sin 6 S . 

arcs where a seismic plane is well de- sin a 2 = — — sin ai (1) 

fined (Fig. 85), the conversion region v 8 smO P 

lies on or near that plane. Figure 87 where 6 S and 6 P are the angles with the 










"1 — 
1 1 





1000 KM 

iTi 266B2421 13 21 2 154 2131 TT~~ 
136933 1322 1 3112213 1 1 
11 1 11 234422211163 31353 3552511 2 1 
21 1 8996 1 1 2141 269 12255531 1 
92554225^56263432231 2422231 53B1 2 122 
/ 44341 143323 1211 2 311522 111111142 
^221 ED95 223 31 53123 4 11112122112312211 
/ 5522 21111 1 122 1 1 

'11 35417311112111 11 1 21 1 1 

' 1 <-• 1 

///1 3*2 1 
#3'1 1 
1,2) '\ 3311 

1 111 



1 1 

25 sec 

30 °U--~ 


Fig. 86. Vertical cross section of the seismicity (U.S. Coast and Geodetic Survey and 
National Ocean ographic and Atmospheric Administration, 1964-1971) for earthquakes inside 
the inset on the map. For the station KMU (Hokkaido, Japan) the ScS-ScSp time using J-B 
travel times is indicated by the dashed lines, and the contour representing the observed time 
difference of eight seconds is hatched. The circle represents the preferred conversion region for 
ScS to ScSp based on amplitude studies. 

vertical of the incident S and the out- 
going P, respectively (Fig. 88). For the 
5 March 1965 event, O s is - 8°. Taking 
a, ~ 80°, P ~ 60° (which is near graz- 
ing for a 30° dipping interface) and 
Vp/vs ~ 1.7, we obtain a value for « 2 of 
~ 16°. Using the preferred conversion 
region (from Fig. 87), the approximate 
projected conversion region is at A in 
Fig. 89, and the projected region based 
on Equation 1 is at B. A and B are 
about 40 km apart. Hence, assuming 
that the interface is laterally homoge- 
neous over such distances near the sta- 
tion, the approximation is a good one. 

The anomalous high-frequency arri- 
vals. Anomalous high-frequency arrivals 
have been observed at Peruvian and 
Chilean stations from deep-focus South 
American earthquakes (Isacks and Bara- 
zangi, 1973; Snoke et al, 1973, 1974a, 
19746, 1974c). Similar arrivals have 
also been observed in Japan (Tsujiura, 
1972) and Tonga (Oliver and Isacks, 
1967; Barazangi et al, 1972). These ar- 
rivals are shear waves with a greater 

travel time, higher frequency content, 
longer duration, and lower apparent 
velocity than direct S. Examples of such 
arrivals (labeled AHFA) are shown in 
Fig. 90. Figure 89 shows the stations and 
location of events for which this arrival 
has been observed. The more northern 
(Peru-Brazil) events were observed at 
ANT, northern Chile, and the southern 
events (Argentina) at the other stations. 
Two models have been proposed which 
can explain these arrivals: the wave- 
guide model (Isacks and Barazangi, 
1973) and the reflection model (Snoke 
et al, 1973, 1974a, 19746, 1974c). These 
models were compared in Year Book 73 
(Snoke et al, 19746), and it was con- 
cluded that although the reflection model 
may not be unique, it provides an ade- 
quate explanation for the arrivals, in- 
cluding the frequency content. It is pre- 
ferred to the waveguide model because 
it makes fewer unverified assumptions, 
and it predicts the observed long- 
distance cut-off (see Fig. 117 in Snoke, 
19746). The following discussion per- 



tains to the reflection model. Use of this sion region, the location of the earth- 
phase to locate an interface is much less quake and the direction of the initial S 
precise than the use of ScSp, because it or P phase. 

is much more sensitive to the velocity- The main tectonic implication of these 

depth model used, the detailed structure arrivals is the existence of a dipping 

of the dipping interface in the conver- interface in the depth range 80-250 km 





200 NNA 
_l I L 








1000 KM 


1 1 1 11 1131 1 1 

> 1112122 1 1 1111 1222141444 1 

"24351443396462534321 21 12 2 1221 4 412247/JC71321 11 1 

12132235C3331 1 v 1 11 2 221 1421 1 

» 1 12 2444513 21 ^ N 1 \ i 11 12 11 1 

1 1444622111 1 s \ 1 1 11 

I 1311 1 N \ 1 11 
1 21 111 \ < 11 1 11 

II 141 \ 3 \ 1 1 1 11311 1 1 
Tlv. 1 1 1x3 M 11 1 11 1 1 1 

11*%. 1 \ \ 1 1 113 2 11 

%l \ \1 n 1 mi 21 1 

1 \. 1 \ 1 V 1 11 222 

1 1 1 






-^ 1 


21 1 

\ 1 11 1 

V\\i\ 1 

\ \ \ \ 

20sec15 10 5 

12 12311 
31434 31 

1 1 



1000 KM 

1965 Mar. 5 SANTIAGO DEL 

m = 5.6 H = 555 KM ESTERO 

A = 19.6° AzB -- 141.2° 




20 sec 

Fig. 87. Vertical seismicity cross section and traced seismograms of ScS and ScSp arrivals 
at NNA (central Peru) ; the circle represents preferred conversion region. Isacks and Molnar's 
(1971) gently dipping slab model is shown in the inset at the lower left. Their seismicity cross 
section is taken over the same region as the one above it, but they conclude onlv M > 5 
(USCGS, 1961-1967). 



Fig. 88. Schematic showing two projects of the geometry for 
the ScS-io-P conversion at the dipping interface. 



80" W 

Fig. 89. Location map showing seismograph stations and epicenters (solid triangles) of 
earthquakes for which the anomalous high-frequency arrivals have been observed. Included 
are projections for the seismic activity for depths of 100-250 km (numbers) and depths greater 
than 500 km (open triangles). Also shown are ray paths for ScSp and the anomalous high 
frequency arrival from the circled event. 

which represents a sufficiently sharp 
velocity discontinuity to support efficient 
reflection of P or S phases incident from 
below. A schematic of the reflection 
model is shown in Fig. 91: The high- 
frequency anomalous arrivals result 
from an S-to-S or a P-to-S reflection at 
a dipping interface. The observed dis- 
tances and arrival time onsets are con- 
sistent with the model calculations based 
on S-to-S reflections where the reflec- 
tions can occur anywhere in a region 
between depths of 80 and 250 km and of 
similar lateral extent. If one assumes 
that the interface is the upper boundary 
of the subducting plate and that it must 

therefore be correlated with the seis- 
micity of ScSp, the possible location of 
conversion regions is decreased consid- 
erably. (The 10-20 second duration of 
the arrival indicates that the conversion 
occurs over a much larger region than 
does ScSp.) For all cases studied, the 
initial phase is necessarily an S phase. 
An application of the two techniques 
for a single- station earthquake pair. 
Both an ScSp and an anomalous high- 
frequency arrival were observed at NNA 
from a deep-focus event on 5 March 
1965. The epicenter for this event is cir- 
cled in Fig. 89. The projected path for 
the ScSp arrival was discussed above, 




-WW - TMfal^J^ 



ARE 10 AUGUST 1967 
h = 550 KM A =9.8° 

h = 599 KM A =13.6° 


NNA 5 MARCH 1965 
h = 555 KM A = 19.6° 


10 20 

30 sec 

Fig. 90. Tracings of short-period records from ARE and NNA showing anomalous high- 
frequency arrivals (labeled AHFA) and the corresponding S arrivals for various epicentral 
distances. The arrival labeled X is interpreted as an AHFA-phase-to-P conversion at the 







O 400 


600 L 



Fig. 91. A schematic of the reflection model. The actual path will be along A a and A 2 rather 
than the straight-line path A. Given A, velocity-depth model(s), a focal depth, and a con- 
version depth, and assuming Snell's Law at the conversion, one can calculate travel time, dip 
angle (0), and strike angle (/3) for each choice of phases 1 and 2. 




400 600 





i i i 

1 1 2 

] 21123 21 1 
1 11 

1 1 

800 1000 

"~l 1 1 — 


12 2 1 1 

1 1 

1 1 1 11 111 2 

1 11 1 21231 31 2 

1 345514321 122 1 1 
1 24264594344 21 
11 5135965551 
1 1 21 12133 
1 12 1 

2 11 11 2 
1123 42 
2 1 1 

• o° 


nnaV.^v.^ v 


^ — i 


ANTjpKjf? "• 


30° S 



1 o 

f o 


1 11 

1 11 

113 3 
2 1 23 22 

Fig. 92. Vertical seismicity cross section for events inside the inset. The circle indicates the 
conversion region for the S-to-S reflection leading to the observed anomalous high-frequency 
arrival at NNA from the event on 5 March 1965. 

and a tracing of the seismic arrival is 
given in Fig. 87. Tracings of two com- 
ponents of the anomalous arrival are 
included in Fig. 90. Assuming a 4% 
higher velocity in the subducting plate 
than in the normal mantle at the same 
depth, one finds a possible path whose 
projections are shown in Figs. 89 and 92. 
The conversion region is denoted by C 
in Fig. 89 and a circle in Fig. 92. 

One sees from Fig. 89 that although 
the ScSp and anomalous high-frequency 
phases sample the interface at roughly 
the same depth, the conversion points 
are about 1400 km apart. Hence, even 
though the earthquake and the station 
are the same, the two phases sample dif- 
ferent parts of the plate boundary. 

Concluding remarks. The conversion 
region for ScSp at a given station has a 
fixed location which, to a good approxi- 
mation, is independent of the location of 
the earthquake sources. Hence, using 
ScSp observations at a single station 
provides information about the location 
of only a small part of the dipping inter- 

face. To map large regions of a plate 
boundary using ScSp therefore requires 
a large number of appropriately located 

The conversion region giving rise to 
the high-frequency phases depends on 
the locations of the station and the 
earthquake relative to the interface, and 
its location is defined much less precisely 
than is the ScSp conversion region. 
These arrivals are useful in that their 
presence indicates the existence of a 
dipping interface, or, if the plate-bound- 
ary is already fairly well located by 
other means, it may provide information 
about the velocity contrasts which can- 
not be obtained from ScSp or seismicity. 


Barazangi, M., B. Isacks, and J. Oliver, 
Propagation of seismic waves through 
and beneath the lithosphere that de- 
scends under the Tonga island arc, J. 
Geophys. Res., 77, 952-958, 1972. 

Isacks, B. L., and M. Barazangi, High- 
frequency shear waves guided by a 


continuous lithosphere descending be- which the radiation and deformation from 

neath South America, Geophys. J. R. earthquakes are interpreted in terms of 

Astron. Soc, 33, 129-139, 1973. theoretical source models should enable 

Okada, H., Forerunners of ScS wave one to estimate the earthquake's geo- 

from nearby deep earthquakes and metric parameters (fault length and ori- 

upper mantle structure in Hokkaido entation) and the physical parameters of 

(in Japanese) Jishin, Ser. 2, 24, 228- the earthquake's environment (strength 

239, 1971. of rock and tectonic stress field). 

Okada, H., Converted P phase from the For the past fifteen years the most 

ScS phase at the inclined deep seismic commonly used source model has been 

zone, in Carnegie Inst. Wash. Year the dislocation model, which assumes 

Book 72, pp. 238-245, 1973. some specific time history for the dis- 

Okada, H., Geophysical implications of placement across the fault. Such an ap- 

the phase ScSp on the dipping litho- proach seemed to us to assume informa- 

sphere underthrusting western South tion that should more properly be derived 

America, in Carnegie Inst. Wash, from the model. Application of the dis- 

Year Book 73, pp. 1032-1039, 1974. location model to date has involved 

Okada, H., Ph.D. thesis, Hokkaido Uni- highly simplified and unsatisfactory 
versity, Sapporo, Japan, in prepara- representations of an earthquake rup- 
tion, 1975. ture. We therefore sought a more real- 
Oliver, J., and B. Isacks, Deep earth- istic earthquake source model. To that 
quake zones, anomalous structure in end, we investigated a version of a stress- 
the upper mantle and the lithosphere, relaxation model that Archambeau had 
J. Geophys. Res., 72, 4259-4275, 1967. introduced and solved. In this model cer- 

Snoke, J. A., I. S. Sacks, and H. Okada, tain details of the rupture process and 

Empirical models for anomalous high- the reaction of the medium are specified, 

frequency arrivals from deep-focus and Archambeau's solution treated the 

earthquakes in South America, in problem as an initial-value problem. 

Carnegie Inst. Wash. Year Book 72, Archambeau's papers were quite contro- 

pp. 233-238, 1973; ibid., Geophys. J. versial because his solution for the radi- 

R. Astron. Soc, 37, 133-139, 1974a. ation-zone displacement amplitude had 

Snoke, J. A., I. S. Sacks, and H. Okada, spectral peaks that did not exist in other 

A model not requiring continuous solutions for apparently similar models, 

lithosphere for anomalous high-fre- For the class of earthquakes we were 

quency arrivals from deep-focus South studying, Archambeau's spectral struc- 

American earthquakes, in Carnegie ture was not inconsistent with the data. 

Inst. Wash. Year Book 73, pp. 1039- A closer examination of the underlying 

1046, 19746; ibid., Phys. Earth Planet, assumptions in Archambeau's approach 

Interiors, 9, 199-206, 1974c. has shown, however, that the stress- 

Tsujiura, M., Spectra of body waves and relaxation model is equivalent to a dis- 

their dependence on source depth: 1. location model and that Archambeau's 

Japanese arc, J. Phys. Earth, 20, 251- spectral peaks resulted from physically 

266, 1972. inadmissible applications of his method 

of solution (see Snoke, Archambeau's 

Seismic Source Studies elastodynamical source-model determi- 

J. A. Snoke, A. T. Linde, and I. S. Sacks nation, this Report) . 

Hence, for a rupture process that oc- 
curs instantaneously, the stress-relaxa- 

General Introduction 

Elastic strain, which is built up slowly tion models and the dislocation models 
in the earth due to various processes, can are equivalent. The source-model solu- 
be released by earthquakes. Studies in tions for a finite rupture velocity are not 



exact (see Snoke, The implications of 
moving boundaries in elastodynamics, 
this Report), but we believe that the 
radiation predicted by a model with a 
finite, smoothly varying rupture velocity 
will not differ significantly from that for 
the instantaneous-rupture models. There 
is some evidence that at least some earth- 
quakes are multiple ruptures. A pre- 
liminary study (see Linde et al., Multiple 
rupture earthquakes and the determina- 
tion of source parameters, this Report) 
indicates that multiple ruptures produce 
spectra which, if interpreted in terms of 
a single-rupture model, lead to erroneous 
estimates for source parameters such as 
fault length and stress drop. 

Our theoretical studies aimed at de- 
veloping more realistic source models are 
continuing together with observational 
studies on selected suites of earthquakes. 

Archambeau's Elastodynamical 
Source-Model Solution 

J. Arthur Snoke 


Earthquake source models provide a 
means by which the radiation and de- 
formations due to earthquakes can be 
interpreted to provide information about 
the properties of the earth in the vicinity 
of the earthquake. In recent years there 
have been a number of different earth- 
quake-rupture models proposed — none 
of them very realistic physically — and 
there have been an almost equal number 
of mathematical approaches introduced 
to solve these models. The most com- 
monly used model at present is that of 
Brune (1970) which is the model derived 
by considering the effective stress avail- 
able to accelerate the sides of the fault. 
Brune's solution is empirical, but its 
near-field and far-field displacements for 
*S-wave radiation are in reasonable agree- 
ment with those derived from dislocation 
models (e.g., Haskell, 1964). 

Another type of model is the stress- 
relaxation model whereby the rupture 
process consists of a phase change in a 

prescribed region resulting in the release 
of a tectonic prestress. The instantaneous 
creation of a sphere has been solved 
exactly by Hirasawa and Sato (1963a, 
19636) for a spherical cavity and by 
Koyama et al. (1973) for a noncavity, 
zero-rigidity sphere. Archambeau (1968, 
1972) has also treated this problem, but 
his solution in general is not exact be- 
cause it neglects to take into account 
dynamical boundary conditions on the 
surface of the rupture shell. Randall 
(1966) has introduced a variant on this 
model in which the postrupture medium 
has no material boundaries. For this 
case, Archambeau's method for obtain- 
ing a solution is exact. 

Although Archambeau's method of 
solution is not exact for most problems 
to which it has been applied, it is still 
potentially useful because it is much 
simpler analytically than the exact solu- 
tion and because it is easily generalized 
to models for non-instantaneous rup- 
tures. Before it can be applied in these 
cases, however, it is necessary to esti- 
mate the validity and closeness of the 

Spectral Peaks 

There has been considerable contro- 
versy regarding Archambeau's method 
of solution because two models employ- 
ing it are the only earthquake source 
models at present which predict peaks 
in the long-period part of the far-field- 
displacement amplitude spectrum with- 
out explicitly assuming rebound or mul- 
tiple ruptures along the fault plane. 
These models have generated interest 
because some long-period amplitude 
spectra of earthquakes seem to exhibit 
definite spectral peaks (Linde and Sacks, 
1971; Rodriquez and Kisslinger, 1972; 
Archambeau, 1974), or, equivalently, 
there is some evidence for rebound in 
the time-domain displacement (Bol- 
linger, 1970; Burdick and Helmberger, 

In his early papers (Archambeau, 
1968, 1972) Archambeau's displacement 



amplitudes had pronounced low-fre- 
quency spectral peaks. Randall (1973; 
see also Snoke et al., 1973) showed that 
these peaks were spurious. However, 
two recent applications of Archambeau's 
method of solution have led to spectral 
peaks: One model is Archambeau's ear- 
lier peak-producing model with a new 
interpretation (Minster, 1973; Archam- 
beau and Minster, 1974; Archambeau, 
1974) , and the other is a different model 
(Snoke, 1974). Spectral peaks in these 
models have been interpreted as result- 
ing from the assumption of a nonuniform 
prestress. There is general agreement 
that no long-period spectral peaks result 
if a uniform prestress is assumed. 

Molnar et al. (1973) and Brune (1974, 
private communication) conjecture that 
the existence of the spectral peaks in 

models using Archambeau's method of 
solution is a consequence of the assump- 
tion that the rupture-zone boundary be- 
comes a free surface following the onset 
of rupture. Molnar et al. assume that 
for earthquakes, friction on the fault 
surface is sufficient to prevent rebound. 
If they are correct, a maximum in the 
amplitude spectrum should occur at zero 

Brune argues that the radiation could 
not be affected by prestress nonuniformi- 
ties if it results entirely from motion on 
the rupture-zone boundary. He notes 
that for the nonuniformity to affect the 
radiation, energy which leaves the fault 
surface at the onset of the rupture must 
have time to sample the nonuniform re- 
gion and be reflected back to the fault 
surface before the fault locks. This is 




t 0. 






Stress — Relaxation 

(X=/i for r >R ) 

Exact Solutions 

( K continuous 
— S(££=O.II Hz) 
- ----P(I/ c =0.20Hz) 

Transparent- Source Model 
(envelope only 

S(Z^=0.I6 Hz 

P(Z/ c =0.28Hz) 





Fig. 93. Far-field displacement amplitude spectra for P and S radiation for the exact solu- 
tions and Archambeau's transparent-source model solutions of the stress-relaxation model. 
For all cases A = n outside the rupture zone, and for the exact solutions the compressibility 
is continuous across the rupture zone boundary. The radius of the rupture zone, Ro is 10 km 
and v p = 10 km/sec. Only the envelope for Archambeau's solutions is shown. 



improbable because the prestress is taken 
to be uniform over a region that is large 
compared to the rupture zone. Brune 
speculates that spectral peaks might re- 
sult from energy reflected from the 
boundary of the uniformly prestressed 
region, causing fault rebound along the 
still-free rupture zone. Calculations show, 
however, that the effects of reflected 
waves are an order of magnitude too 
small to account for the spectral struc- 
ture predicted in Snoke's model (Snoke, 

Results and Conclusions 

This study was undertaken to resolve 
this dilemma and to determine the 
validity of the application of Archam- 
beau's method when it does not lead to 
an exact solution. Given below is a sum- 
mary of the conclusion of that study 
(see Snoke, 1975, for the more detailed 
account) . 

1. Contrary to a claim made by Ar- 

chambeau (1972), there is no simple re- 
lationship between the ratio of P- and 
S-radiated energies and the ratio of the 
energy released from the hydrostatic and 
deviatoric parts of the prestress. 

2. Archambeau's (or, equivalently, 
Randall's 1964) elastodynamical source- 
model solution for an instantaneous rup- 
ture is equivalent to an instantaneous 
stress pulse and/or dislocation on the 
rupture-zone boundary. (See also Min- 
ster, 1973.) 

3. Figure 93 shows the frequency do- 
main amplitude for Archambeau's solu- 
tion and an exact solution for the P-wave 
and $-wave radiation resulting from the 
instantaneous creation of a zero-rigidity 
sphere in a uniformly (shear) pre- 
stressed, elastically homogeneous me- 
dium. Archambeau's solution is seen to 
be a good approximation to the exact 
solution for these cases: It has the same 
w -2 high-frequency fall-off and zero- 
frequency limit in the frequency domain. 






Archambeau s solution 

R D model 

R D =I0R /t 2 =O.I/i, 
R s = oo 

R s model 

R s =IOR /i 2 =/a, 

//^//Acausal behavior 


[vpt-(r-R j]/R c 

Fig. 94. Time-domain solutions for the P-wave far-field displacement in arbitrary units for 
the R s model (Minster, 1973) with infinite and finite Rs and for Snoke's (1974) R D model 
solved by Archambeau's method, (/x and /ai are the outside and inside rigidities, respectively, 
in the R D model.) R = 10 km and v p = 10 km/sec in all cases. The acausal parts of the 
solutions are hatched. 










100 — 





R S = I0R 

Nonuniform prestress 

R ,R D and/i| fixed 

fi 2 =X/i, 





Fig. 95. P-wave far-field displacement ampli- 
tude spectra for the Rd model solved by 
Archambeau's method with R = 10 km, Rd = 
100 km, v p = 10 km/sec and the inside rigidity 
fix held constant while the outside rigidity tx is 
varied. The spectra for the (nonexact) causal 
solution for models are approximately equal to 
the non-discontinuity (fii — \x) spectrum for 
all values of fi. The dashed line represents an 
Rs model spectrum with Rs = 10 Ro = 100 
km. The amplitude scale is in arbitrary units. 

In the time domain, Archambeau's solu- 
tion predicts a one-sided parabolic pulse 
which is quite similar in shape to the 
exact solution. 

4. From the stress-pulse/dislocation- 
model representation for Archambeau's 
solution it follows that all discontinuity 
surfaces in the elastic medium radiate 
simultaneously for an instantaneous rup- 
ture. Physically, only the rupture zone 
would be a source of radiation, so that 
Archambeau's prescription for the solu- 
tion cannot be applied directly when 
there are discontinuities in the medium 
other than that of the rupture-zone 
boundary. Two models, the Rs model 
(Minster, 1973) and the R D model 

(Snoke, 1974), have additional discon- 
tinuities: that direct application of 
Archambeau's prescription for the solu- 
tion results in peaks in the long-period 
part of the far-field displacement ampli- 
tude spectrum is due only to assumptions 
that are physically impossible (see Figs. 
94 and 95). 

The controversy concerning the spec- 
tral peaks has therefore been resolved: 
all spectral peaks produced by models 
employing Archambeau's method of so- 
lution are spurious. Hence, observed 
spectral structure cannot be explained by 
any instantaneous rupture model using 
Archambeau's method of solution. 


Archambeau, C. B., General theory of 
elastodynamic source fields, Rev. Geo- 
phys., 16, 241-288, 1968. 

Archambeau, C. B., The theory of stress 
wave radiation from explosions in pre- 
stressed media, Geophys. J. R. Astron. 
Soc, 29, 329-366, 1972; Appendix, 31, 
361-363, 1973. 

Archambeau, C. B., Investigations of 
tectonic stress, Semiannual Technical 
Report No. 1, Contract No. F19628- 
74-C-0087 ARPA/AFCRL, Univer- 
sity of Colorado, 1974. 

Archambeau, C. B., and J. B. Minster, 
Elastodynamic representation theorem 
and the equivalence of various source 
theories, Trans. Am. Geophys. Union, 
55, 683, 1974. 

Bollinger, G. A., Fault length and frac- 
ture velocity for the Kyushu, Japan 
earthquake of October 3, 1963, J. Geo- 
phys. Res., 75, 955-964, 1970. 

Brune, J. N., Tectonic stress and the 
spectra of seismic shear waves from 
earthquakes, J. Geophys. Res., 75, 
4997-5009, 1970; erratum, 76, 5002, 

Burdick, L. J., and D. V. Helmberger, 
Time functions appropriate for deep 
earthquakes, Bull. Seismol. Soc. Am., 
64, 1419-1428, 1974. 



Haskell, N. A., Total energy and energy 
spectral density of elastic wave radi- 
ation from propagating faults, Bull. 
Seismol. Soc. Am., 54, 1811-1841, 1964. 

Hirasawa, T., and R. Sato, Propagation 
of elastic waves from a spherical ori- 
gin: Part 1 and Part 2 (in Japanese), 
Zisin, 16, 52-77, 1963a,6. 

Koyama, J., S. E. Horiuchi, and T. Hira- 
sawa, Elastic waves generated from 
sudden vanishing of rigidity in a 
spherical region, J. Phys. Earth, 21, 
213-226, 1973. 

Linde, A. T., and I. S. Sacks, Errors in 
the spectral analysis of long-period 
seismic body waves, J. Geophys. Res., 
76, 3326-3336, 1971. 

Minster, J. B., Elastodynamics of failure 
in a continuum, Ph.D. thesis, Cali- 
fornia Institute of Technology, 1973. 

Molnar, P., K. H. Jacob, and K. Mc- 
Camy, Implications of Archambeau's 
earthquake source theory for slip on 
faults, Bull. Seismol. Soc. Am., 63, 
101-104, 1973. 

Randall, M. J., On the mechanism of 
earthquakes, Bull. Seismol. Soc. Am., 
54, 1283-1289, 1964. 

Randall, M. J., Seismic radiation from 
a sudden phase transition, J. Geophys. 
Res., 71, 5297-5302, 1966. 

Randall, M. J., Spectral peaks and earth- 
quake source dimensions, J. Geophys. 
Res., 78, 2609-2611, 1973. 

Rodriguez, R., and C. Kisslinger, Spec- 
tral classification of New Madrid 
earthquakes, Trans. Am. Geophys. 
Union, 53, 450, 1972. 

Snoke, J. A., Earthquake radiation spec- 
trum from models with concentrated 
prestress, in Carnegie Inst. Wash. 
Year Book 73, pp. 1048-1054, 1974. 

Snoke, J. A., Archambeau's elastody- 
namical source-model solution and 
low frequency spectral peaks in the 
far-field displacement amplitude, Geo- 
phys. J. R. Astron. Soc, in press, 1975. 

Snoke, J. A., A. T. Linde, and I. S. Sacks, 
Source-spectral relations for earth- 
quakes, in Carnegie Inst. Wash. Year 
Book 72, pp. 245-246, 1973. 

The Implications of Moving 
Boundaries in Elastodynamics 

J. A. Snoke 


The preceding report (Snoke, 1975a; 
see also Snoke, 19756) contains a com- 
parison of Archambeau's (1968, 1972) 
solution with the exact solution for the 
radiation-zone displacement from the in- 
stantaneous creation of a zero-rigidity 
sphere of radius R in a homogeneous 
medium with a uniform (shear) pre- 
stress. The creation of zero rigidity is 
brought about by a phase transition in- 
volving no external forces. Archambeau 
(1968, 1972) also treats the generaliza- 
tion of this model whereby the sphere 
grows with a finite (constant) rupture 
velocity V for the time interval < t < 
T = R /V. As for the instantaneous 
case, Archambeau's solution is not exact, 
but Burridge (1975) has derived what 
he claims to be an exact solution for this 
problem. In this report it is shown that 
Burridge's solution also is not exact be- 
cause he has not taken into account the 
effect of the motion of the rupture-zone 
boundary on the boundary conditions. 

Archambeau's Solution 

For the instantaneous creation of a 
zero-rigidity sphere of radius R in a 
homogeneous, uniformly prestressed me- 
dium, the ;th component of the displace- 
ment is taken to be given by Uj (r, t; R). 
Using superposition and convolution, 
Randall (1973) finds that the displace- 
ment Uj(r, t) for the finite-rupture- 
velocity case is given by 

Uj (r,t) -[ drVdUj 

^ o 

\?,t-r;R (t)]/3B (1) 

where V = dR/d T for < r < T. One 
can derive Archambeau's (1968, 1972) 
form of the solution from Equation 1. 
(Randall also gives an expression for the 
frequency-domain displacement in his 
Equation 9, but this expression is incor- 



rect because its derivation neglects the 
implicit dependence on Uj of r through 

For the Uj to be exact solutions, they 
and their spatial derivatives must satisfy 
certain boundary conditions at r = R(t) 
for all times. For a given set of Uj, any 
boundary conditions to be satisfied by 
the Uj as defined by Equation 1 are de- 
termined through that equation. Archam- 
beau's Uj are not exact because they 
satisfy the correct boundary conditions 
only in the infinite-time limit — they 
satisfy no dynamical boundary condi- 
tions. The Uj defined by Equation 1 are 
also not exact whether one uses Archam- 
beau's Uj or the exact Uj (Snoke, 1975 b) 
in the right-hand side of the equation. 

Burridge's Solution 

Burridge's solution (1975) for the 
finite, constant rupture-velocity model 
uses an approach developed by Burridge 
and Alterman (1972) which is, in princi- 
ple, exact. Archambeau's solution is in- 
dependent of the compressibility or 
density for r < R, but Burridge's solu- 
tion is for the specific case of zero com- 
pressibility and density — a cavity for 
r < R. 

Burridge's solution is based on the 
boundary condition 

(TijXj\ r — R — 

— <T {0) -X -I 



where the a^hj are the elements of the 
initial prestress tensor; the ay are the 
elements of the stress tensor defined rela- 
tive to the <T {{)) ij', and repeated indices 
are summed from one to three. 

Boundary Conditions for 
Moving Boundaries 

Derivation. The equation of motion in 
elastodynamics is 

d j(Ti j = — Fi 
+ d(pdui/*t)/dt 

where F is the body-force density, p the 
density, and djA = 3A/dz,-. Equation 3 
is valid in an inertial frame, which we 
take to be the observer's frame of refer- 
ence. The problem of interest is what 
Muskhelishvili (1963, p. 79) calls the 
"first fundamental problem;" namely, to 
find solutions to Equation 3 satisfying 
F{r,t) -> f(r,t) on some surface S and 
having the initial conditions u(r,0) = 
u°(r) and ^u{r,t) fdt\ t =o = w(r), where 
7, u° and W are known. 

In the problem of interest, the surface 
S — the surface of the rupture sphere — 
is moving for a period of time in the 
observer's frame of reference. The effects 
of moving boundaries on the boundary 
conditions satisfied by the variables of 
interest have been treated extensively in 
electrodynamics (e.g., the discussion of 
Faraday's law by Jackson, 1962, Ch. 6) 
and hydrodynamics (e.g., the discussion 
of Euler's equation by Landau and 
Lifshitz, 1959, Ch. 1). Minster (1973) 
considers moving boundaries in the con- 
text of Green's-function integral repre- 
sentations in elastodynamics. 

The problem at hand is to reexpress 
the boundary conditions in terms of Uj 
and Gij defined at a given time in the 
observer's frame of reference. This can 
be done using Equation 3 after it has 
been transformed to a reference frame 
in which the surface S is at rest. 

In a coordinate frame moving with 
respect to a fixed frame, the time deriva- 
tive of a function A (f,t) in the moving 
frame must take into account the time 
dependence of r. Designating the trans- 
formation velocity as v = ^f/^t, one has 

^A/'dt\ moxing frame = Da/Dt 

= 3A/3£| fiX ed frame + ^'V^ 


Using Equations 3 and 4, Equation 3 
in the moving frame can be written: 

"djtr'ij = — F\ 
+ D( P Dui/Dt)/Dt 


(3) where 





+ D{ P UiVj)/Dt 
+ pVj(Dui/Dt 

— VrfinUi 
+ ufinVn) 


are the elements of the stress tensor in 
the moving frame and 

F\ = Fi 

- Diupjipvdym 

— p(d n v n )Dui/Dt 
— ufij( P vjd n v n ) (7) 

are the components of the body force 
in that frame. 

The boundary condition can be derived 
from Equation 5 (e.g., Landau and 
Lifshitz, 1959, Ch. 1). The result is 

\_(T f ij7lj] /'.j 


where n(r) is the unit normal to the 
surface S at r on.S (defined so that n'r 
> 0), [A] represents the jump discon- 
tinuity in A across S and — jr is the 
pressure on S resulting from any singular 
terms on the right-hand side of Equation 
5 at f and t on S. 

Equations 5-8 differ from Minster's 
results (Minster, 1973, p. 54) because 
Minster has incorrectly transformed only 
one of the two time derivatives on the 
right-hand side of Equation 3 into the 
moving frame. 

Applications. In the problem of in- 
terest, the non-instantaneous creation of 
a spherical cavity in a uniformly pre- 
stressed medium, the surface S is given 
by r = R(t) =VtH(T — t)H(t), where 
H is the Heaviside function, so that the 
velocity v is given by v = VH{T — t) u 
Hit)?. The initial values are ^° = w = 
0, and F = / = as well. Because the 
rupture sphere is taken to be a cavity, 
(Ty = —cr m ij for r < R. Using p = 
p <>H(r — R) one finds that /' = 0, and 
Equation 8 takes the form: 


Ijaj r 





= [- 

— P °V(2Dui/Dt 

+ 3Vu i )H(T-t)']H(t) 




Hence, Burridge's boundary condition 
(Equation 2, above) is not correct for 
< t < T. 

A commonly used model for an earth- 
quake is a planar rupture zone (e.g., 
Brune, 1970). For a non-instantaneous 
rupture, n is perpendicular to the plane 
while v lies in the plane so that v • n = 
0. Hence, from Equation 6, o-'^% = 
(TijUj on each side of S so that the bound- 
ary conditions take the same form as for 
a stationary boundary. 

Concluding Remarks 

It has been shown that Burridge's 
solution (Burridge, 1975) is not exact 
for the problem of a growing spherical 
cavity in a uniformly prestressed me- 
dium because it assumes incorrect bound- 
ary conditions during the time the sphere 
is growing. The approach developed by 
Burridge and Alterman (1972) could in 
principle provide the exact solution for 
the problem if the correct boundary con- 
dition, Equation 9, were used. However, 
Burridge (1975) states that this ap- 
proach for deriving exact solutions is 
limited in application to a class of closely 
related problems, and a preliminary in- 
vestigation indicates that replacing 
Equation 2 by Equation 9 takes the 
problem of interest out of that class. 

The lack of an exact solution for this 
problem makes it difficult to choose be- 
tween Burridge's and Archambeau's 
(1972) solution as to relative correctness. 
Such a choice is only of academic inter- 
est because the two solutions produce 
quite similar far-field displacements in 
both the time and frequency domains, 
provided one uses the corrected form of 
Archambeau's solution (Randall, 1973; 
Snoke et al. } 1973; Snoke, 1975a, 19756). 
Based on comparisons between the exact 
solution and Archambeau's solution for 
an instantaneous rupture (see the pre- 
ceding paper and Snoke, 19756), one 



would anticipate that these solutions 
differ little from the exact solution. 


Archambeau, C. B., General theory of 
elastodynamic source fields, Rev . Geo- 
phys., 16, 241-288, 1968. 

Archambeau, C. B., The theory of stress 
wave radiation from explosions in pre- 
stressed media, Geophys. J. R. Astron. 
Soc, 29, 329-366, 1972; 31, 361-363, 

Brune, J. N., Tectonic stress and the 
spectra of seismic shear waves from 
earthquakes, J. Geophys. Res., 75, 
4997-5009, 1970; erratum, 76, 5002, 

Burridge, R., The pulse shape and spec- 
tra of elastic waves generated when a 
cavity expands in an initial shear field, 
J. Geophys. Res., 80, 2606-2607, 1975. 

Burridge, R., and Z. Alterman, The elas- 
tic radiation from an expanding spheri- 
cal cavity, Geophys. J. R. Astron. 
Soc, 30, 451-477, 1972. 

Jackson, J. D., Classical Electrodynam- 
ics, John Wiley and Sons, 1962. 

Landau, L. D., and E. M. Lifshitz, Fluid 
Mechanics, Addison- Wesley, 1959. 

Minster, J. B., Elastodynamics of failure 
in a continuum, Ph.D. thesis, Cali- 
fornia Institute of Technology, 1973. 

Muskhelishvili, N. I., Some Basic Prob- 
lems of the Mathematical Theory of 
Elasticity, P. Noordhoff Ltd., The 
Netherlands, 1963. 

Randall, M. J., Spectral peaks and earth- 
quake source dimensions, J. Geophys. 
Res., 78, 2609-2611, 1973. 

Snoke, J. A., Archambeau's elastody- 
namical source-model solution and 
low-frequency spectral peaks in the 
far-field displacement amplitude, Geo- 
phys. J. R. Astron. Soc, in press, 1975. 

Snoke, J. A., A. T. Linde, and I. S. Sacks, 
Source-spectral relations for earth- 
quakes, in Carnegie Inst. Wash. Year 
Book 72, pp. 245-246, 1973. 

Multiple Rupture Earthquakes and 

the Determination of Source 


Alan T. Linde, I. Selwyn Sacks, 
and J. Arthur Snoke 


A number of theoretical models for 
seismic sources (e.g., Ben-Menahem, 
1962; Haskell, 1964; Aki, 1967; Brune, 
1970; Archambeau, 1968) have yielded 
solutions for the radiation field. Snoke, 
in a companion paper (Archambeau's 
elastodynamical source-model solution, 
this Report) discusses certain of these 
models and concludes that in most im- 
portant respects they do not differ funda- 
mentally from one another. In all models 
which have been proposed (whether dis- 
location or initial value) the problem 
solved is that of stress release due to 
simple rupture in an infinite homogene- 
ous medium subjected to a uniform shear 
field. Numerous studies have been re- 
ported in which the theoretical solutions 
have been used to infer earthquake 
source parameters (dimensions and stress 
release) from observed far-field spectra. 
California earthquakes have been exten- 
sively studied through this approach; 
Fig. 96, from Tucker and Brune (1973), 
is illustrative of the spread in results 
obtained. Spottiswoode and McGarr 
(1975) applied this method to tremors 
in a deep-level gold mine and obtained 
stress drops ranging from 5 to 50 bars. 
However, in situ observations indicate 
localized stress drops of about 2 kbar, 
which agree with laboratory tests. 

Spottiswoode and McGarr also re- 
ported from studies of underground 
damage that many of the tremors are 
multiple events separated in space and 
time. Furthermore, there is an increasing 
body of evidence that large earthquakes 
are multiple events (Wyss and Brune, 
1967; Trifunac and Brune, 1970; Gupta 
et al., 1971) and, in a model study, Brune 
(1973) found multiple events were com- 

Here we present a preliminary investi- 




gation to determine whether or not some (see Snoke, Archambeau's elastodynam- 
of the large variation in reported stress ical source-model solution, this Report) 
drops and the apparently (incorrect) as a reasonable approximation to the 
low stress drops for the mine tremors radiation from a single event and scale 
may be the result of interpreting the the height ratio of the pulses to reflect 
radiation from a multiple source in equal stress drops. We normalize our 
terms of a single-event source model. models by taking the seismic moment 
Source-spectral relationships. In the and total duration of rupture to be con- 
frequency domain, all the theories give stant. If we assume constant rupture 
amplitude spectra which are flat at fre- velocity, this is equivalent to taking the 
quencies lower than a corner frequency, same-sized fault plane with the same 
/o, and decreases above f as a power average slip for all cases, 
of frequency (often /~ 2 ). The models Figure 97 shows the spectrum of a 
relate the physical parameters of the parabolic pulse corresponding to a single 
source to the characteristics of the radi- rupture. The high-frequency asymptote 
ated spectrum. For an earthquake of of the envelope has a slope of / _2 , and 
fault area A with average slip d in a the intersection of this asymptote with 
medium of rigidity /*, the seismic mo- the low-frequency asymptote (zero slope) 
ment M is defines the corner frequency / . Fre- 

fyj- aJ~ q\ quency scales for this and subsequent 

j spectra are plotted in units of / . 

M oc Q (2) Equal duration sub-events. Two pairs 

of equal pulses are shown in Fig. 98. 
where O is the low frequency level of Interference effects at frequencies near / 
the far-field displacement amplitude introduce a change in the shape of the 
spectrum. The corner frequency is re- spectra from that of the single pulse. De- 
lated to the fault dimension by pending on where f lies within the pass- 

f <x 1/L (3) Dan cl of the seismograph, it is possible 

that different corner frequencies could 

If for simplicity we assume complete be picked. 

stress drop, then the stress drop is given The high-frequency asymptote (slope 

by of /~ 2 ) gives a corner frequency /i « 

a oc ad/L (4) 2/ , indicative of the sub-event size 

and hence rather than of the total rupture dimen- 

M oc ctAL (5) sion. If> however, f is high in the seismo- 
graph pass-band, then it is conceivable 

For a fault which grows symmetrically that a lower slope, high-frequency asymp- 

in two dimensions tote wouM be chogen resulting in an 

o- oc M / 3 estimate of f 2 «* l/2/ . For a given seis- 

,„ mic moment we see from Fig. 96, and 

oc 12 f 8 . . 

o-'o Equation 6, that using these different 

(6) values for corner frequency gives stress 

Thus, in principle, determining O and releases differing by a factor of about 60. 

/o from the far-field spectrum allows an There exists a third possible interpre- 

estimate of the stress release. tati <> n ° f the double pulse spectrum: If 

Numerical models of multiple rupture, /o is toward the low-frequency side of 

We limit our discussion to double events the pass-band, the wide interference hole 

(more complex sources show similar ef- in the amplitude spectrum could give the 

fects) and will assume that both sub- appearance of spectral peaks, 

events have the same stress drop. In the N onequal- duration sub-events. For at 

time domain, we use a parabolic pulse least some large shallow earthquakes, 





0.5 1 2 


Fig. 97. A single parabolic pulse and its amplitude spectrum. The asymptotes (dashed line) 
define a corner frequency jo. 




Fig. 98. Equal duration sub-events and their spectra. Asymptotes for the corresponding 
single pulse (see text) meet at jo. The high-frequency asymptote for the double-pulse spectra 
gives a corner frequency /i. The intermediate-slope asymptote defines a corner at jz. 

later sub-events can be larger than the 
initial event (e.g., the Alaskan earth- 
quake of 1964, see Wyss and Brune, 
1967). In Fig. 99 we consider a simple 
example of this case. We compare two 
double ruptures of different duration. 
The pulse-height ratios for cases a and b 
in Fig. 99 reflect equal stress drops for 
the two sub-events for two different 
types of rupture growth. In case a the 
width of the two sub-event fault planes 
is the same, and the pulse height is pro- 
portional to the length (i.e., duration). 
In case b both length and width increase 
(e.g., a circular or elliptical fault), and 
the pulse height is proportional to the 
square of the dimension. 

Over a wide frequency range, an in- 
termediate slope of / _1 exists. Thus, 
depending on the value of / and the 
pass-band of the seismograph, different 
interpretations of the double-event spec- 
tra could be made in terms of a simple 
rupture model. As in the previous exam- 
ple, the calculated stress drops for an 
earthquake of given seismic moment 
could therefore vary considerably. 


We have examined changes in spectral 
behavior as a result of varying the pulse 
lengths and heights and also of varying 
the time between pulses. The spectral 
effects of multiple events discussed above 



hold for a wide range of these parameters 
but tend to become less obvious as one 
pulse is made much larger than the other. 
Similar effects occur if larger numbers 
of pulses are used to represent a more 
complex multiple rupture. 

This simple numerical study indicates 
that the occurrence of multiple ruptures 
may be responsible for several observed 
effects : 

1. The uncertainty and resultant scat- 
ter in calculated stress drops is a conse- 
quence of the fact that the corner fre- 
quency is not a stable indicator of 
rupture dimensions. This is in accord 
with the results presented by Tucker and 
Brune (1973) (see Fig. 96) in that cal- 
culations based on spectra with asymp- 
totes of high slope yield high stress drops 
and vice versa. 

2. Aki (1972) has argued that M s — 
M h observations require a significant fre- 
quency interval with / _1 slope. For 
many of our numerical models, an inter- 
mediate slope of / _1 results. This is par- 
ticularly marked for the models in which, 
for the sub-events, the length of the 
rupture increases but the width remains 

3. If the sub-events are about equal 
in size, then it is possible that the ob- 
served spectra could be interpreted as 
having a peak. Such observations have 
been made (Linde and Sacks, 1972; 

Rodriguez and Kisslinger, 1972; Archam- 
beau and Kisslinger, 1974; Bollinger, 
1970; Burdick and Helmberger, 1974) 
but are not compatible with any of the 
current simple source models, unless re- 
bound occurs on the fault (see Snoke, 
Archambeau's elastodynamical source- 
model solution, this Report). 

Since the corner frequency-seismic 
moment method for determining stress 
drop may result in a wide scatter of re- 
sults (both higher and lower than actual ) , 
we suggest that results should be checked 
independently by using relationships in- 
volving radiated energy. From Aki 
(1966) we have 

o- oc E W /( V M ) 

where the seismic efficiency 

7) — E lc /E s 

the ratio of the seismic wave energy to 
elastic strain energy released. This ap- 
proach is not a simple one, however. 
Elastic strain energy release is difficult 
to estimate for very shallow events, and 
for deep events the measurement is not 
feasible. Moreover, measurement of wave 
energy requires sensing of the radiated 
field over a wide range of frequencies, 
and little is known about seismic effi- 
ciency. Nevertheless, since efficiency has 
an upper bound of 1, it is possible to 
determine a lower limit to the stress 






a / V \ 

/ \ 
/ \ 

/ \ 
' \ 

i V v 

Fig. 99. Nonequal-duration sub-events and their spectra, As in Fig. 98, corners could be 
pulsed at /i and / 2 . The intermediate slope here is /~\ 



drop. (The observed corner frequency 
will depend also on seismic efficiency 
since efficiency is presumably frequency 
dependent. This added complication was 
ignored both in this study and in that 
of Tucker and Brune.) 

The simple models discussed above do 
not appear to explain the large discrep- 
ancies between stress drops determined 
by spectral means and those calculated 
from measurements at the fault (Spottis- 
woode and McGarr, 1975). Nevertheless, 
the results do indicate that the multiple 
rupture model should be considered more 
seriously when attempting to interpret 
far-field spectral observations in terms 
of source parameters. 


Aki, K., Generation and propagation of 
G waves from the Niigata earthquake 
of June 16, 1964, 2, Estimation of 
earthquake moment, released energy, 
and stress-strain drop from the G wave 
spectrum, Bull. Earthquake Res. Inst. 
Tokyo Univ., 44, 73-88, 1966. 

Aki, K., Scaling law of seismic spectrum, 
J. Geophys. Res., 72, 1217, 1967. 

Aki, K., Scaling law of earthquake source 
time- function, Geophys. J. R. Astron. 
Soc, 31, 3-25, 1972. 

Archambeau, C. B., General theory of 
elastodynamic source fields, Rev . Geo- 
phys. Space Phys., 6, 241-288, 1968. 

Archambeau, C. B., and C. Kisslinger, 
Investigations of tectonic stress, Semi- 
annual Technical Report, ARPA Order 
No. 1795, University of Colorado, 1974. 

Ben-Menahem, A., Radiation of seismic 
body waves from a finite moving 
source in the earth, J. Geophys. Res., 
67, 345-350, 1962. 

Bollinger, G. A., Fault length and frac- 
ture velocity for the Kyushu, Japan 
earthquake of October 3, 1963, J. Geo- 
phys. Res., 75, 955-964, 1970. 

Brune, J. N., Tectonic stress and the 
spectra of seismic shear waves from 
earthquakes, J. Geophys. Res., 75, 
4997-5009, 1970. 

Brune, J. N., Earthquake modelling by 
stick-slip along precut surfaces in 
stressed foam rubber, Bull. Seismol. 
Soc. Am., 63, 2105-2119, 1973. 

Burdick, L. J., and D. V. Helmberger, 
Time functions appropriate for deep 
earthquakes, Bull. Seismol. Soc. Am., 
64, 1419-1428, 1974. 

Gupta, H. K., B. K. Rastogi, and H. 
Narain, The Koyma earthquake of 
December 10, 1967: a multiple seismic 
event, Bull. Seismol. Soc. Am., 61, 
167-176, 1971. 

Haskell, N. A., Total energy and energy 
spectral density of elastic wave radi- 
ation from propagating faults, Bull. 
Seismol. Soc. Am., 54, 1811-1841, 1964. 

Linde, A. T., and I. S. Sacks, Dimensions, 
energy, and stress release for South 
American deep earthquakes, J. Geo- 
phys. Res., 77, 1439-1451, 1972. 

Rodriquez, R., and C. Kisslinger, Spec- 
tral classification of New Madrid 
earthquakes, Trans. Am. Geophys. 
Union, 53, 450, 1972. 

Spottiswoode, S. M., and A. McGarr, 
Source parameters of tremors in a 
deep-level gold mine, Bull. Seismol. 
Soc. Am., 65, 93-112, 1975. 

Trifunac, M. D., Stress estimates for the 
San Fernando, California, earthquake 
of February 9, 1971: main event and 
thirteen aftershocks, Bull. Seismol. 
Soc. Am., 62, 721-750, 1972. 

Trifunac, M. D., and J. N. Brune, Com- 
plexity of energy release during the 
Imperial Valley, California, earth- 
quake of 1940, Bull. Seismol. Soc. Am., 
60, 137-160, 1970. 

Tucker, B. E., and J. N. Brune, Seismo- 
grams, S-wave spectra, and source 
parameters for aftershocks of San 
Fernando, California, earthquake of 
February 9, 1971. Geological and Geo- 
physical Studies, Vol. Ill, pp. 69-121, 
U.S. Dept. of Commerce, 1973. 

Wyss, M., and J. N. Brune, The Alaskan 
earthquake of 28 March 1964; a com- 
plex multiple structure, Bull. Seismol. 
Soc. Am., 57, 1017-1023, 1967. 



Borehole Strainmeters : Long-Term 

Stability and Sensitivity to 


/. S. Sacks, J. A. Snoke, Y. Yamagishi, 
and S. Suyehiro 


The performance of the borehole 
strainmeters that are operating in a 
seismically active region near Matsu- 
shiro, Japan, has been evaluated in a 
number of previous studies. The fidelity 
of the instrument when exposed to high 
acceleration such as might be expected 
from strong local earthquakes has been 
tested by Sacks et al. (1971a). The re- 
sponse of the strainmeter and its local 
environment to atmospheric pressure 
fluctuations has been studied by Sacks 
and Evertson (1970) and by Sacks et al. 
(19716). The sensitivity of strainmeters 
to rainfall and pressure changes in the 
aquifer has been determined by Sacks 

Long-Term Stability 

In this report we consider the long- 
term stability of the borehole strain- 

meter and the implications for detec- 
tion of precursory dilatancy. There are 
three strain-measuring instruments op- 
erating at the Matsushiro Seismological 
Observatory. Two borehole volume 
strainmeters, 300 m apart, are situated 
about 250 m from a pair of 100-m-long 
quartz bar horizontal extensometers with 
N-S and E-W orientations. Figure 100 
is a plan showing the locations of the 
various instruments. 

The borehole strainmeters are dila- 
tometers. Using the fact that the exten- 
someters are installed at a free surface, 
the equivalent dilatant (volume) strain 
can be calculated from the data of the 
two horizontal extensometers. The re- 
sult is ^extensometers = [ (1 — 2a) / (1 — <j) ] 

(«ns + «ew) where e N s and e E w are the 
strains measured by the two horizontal 
extensometers and o- is Poisson's ratio. 
Poisson's ratio for the Matsushiro region 
is not known, but must lie in the usual 
range of 0.25 to 0.3. (1 — 2cr)/(l — o-) is 
then in the range 0.66 to 0.57. 

Figure 101 shows the secular strain 
results from the two borehole strainme- 
ters, and the extensometers calculated 
for Poisson's ratio = 0.25. The three 


station )( 

O strainmeter 



IOO meter quartz bar 


no. 2 

Fig. 100. Map of the Matsushiro Seismological Observatory area showing the location of the 
two borehole strainmeters and the 100-m quartz-bar extensometers. 



6 I0~ 6 

v _ 

i — i — r 

i i i i r 

"i — i — r 





o x v x 
x x x Q x x 

° o o 

y Compression 

• 2/3 (e NS +e EW ) 

O Borehole strainmeter no. I 
X Borehole strainmeter no. 2 


X X 

J I L 

6 8 


J I L 



J I I L 



Fig. 101. Secular dilatant strain in the Matsushiro (swarm) region measured by three 
independent instruments located as shown in Fig. 100. 

data sets agree moderately well and in- 
dicate a compressional volume strain 
rate of about 0.4 X 10- 7 /year. The 
borehole instruments give values of 
about 0.36 X 10~ 7 /year and the exten- 
someters give 0.48 X 10~ 7 /year (o- = 
0.25) or 0.41 X 10- 7 /year ( a = 0.3). 
Based on the extensometer data, the 
mean dilatant strain rate since April 
1968 has been 0.66 X 10" 6 per year. 
This strain rate is consistent with geo- 
detic observations which show that the 
horizontal strain in the region of the 
Matsushiro earthquake swarm reached 
3.7 X 10~ 4 (extension) at the swarm's 
peak in October 1966 and has been de- 
creasing since then (Kasahara et al., 

The intrinsic long-term noise of the 
strainmeters can be estimated from how 
well the instruments track. If one cor- 
rects for the differences in secular strain 
(Fig. 101), instrument 2 and the exten- 
someters differ by less than 10~ 7 in the 
monthly readings with a mean difference 
of 0.5 X 10 -7 . The mean difference be- 
tween the monthly readings of borehole 
strainmeter 1 and the extensometers is 
1.3 X 10 -7 . Instrument 1 showed very 
much higher noise than 2 at periods of 
% to 2 hours (Sacks, 1974) , possibly due 

to its proximity to an automatic pump- 
ing station. This noise was assumed to 
be due to fluctuations in the aquifer 
since that has been found to influence 
the local strain field very strongly. 

Response to Dilatancy 

According to the dilatancy model (see 
Nur, 1972), preceding an earthquake 
there is a change in volume in the re- 
gion in which an earthquake will occur. 
It is of interest to know how the bore- 
hole strainmeters (which act as dilatom- 
eters) will react to such slow volume 
changes at some distance from the source 
region. We will approximate the dilat- 
ancy by assuming a point source of ex- 
pansion which is constant in time. (The 
strainmeter is assumed to be outside the 
source region.) The solutions for both 
an infinite space and a half-space can be 
obtained from Cagniard (1962, Ch. 8). 

For an infinite space the displacement 
due to a point source of expansion at 
z = h and p = (in cylindrical coordi- 
nates) is given by 

8V , 

U a = ft3 (PP+ (Z 

h)z) (1) 

where p 2 = x 2 -f- y 2 , R 2 = p 2 + {z — h) 2 
and 87 is the total volume change of any 




Fig. 102. Geometry for a dilatation point source in a half-space bounded by a free surface. 
The source is at z = h and p = 0. 

volume including the source point R = 
(see Fig. 102). The dilatation 6 = v • 
it is zero for R =^= 0. Hence, strainmeters 
outside the source region would not be 
affected by a dilatation source. 

The strainmeters, however, are not 
imbedded in an infinite space but, to a 
good approximation, are at the free sur- 
face of a half-space. Although Cagniard 
has solved this problem, his solution is 
obtained by taking the static limit of a 
dynamic problem. We will outline here 
a more direct derivation. 

The displacement u can be written as 
u = u a -\- u h where u a is given above by 
Equation 1 and S 6 is the solution of the 
homogeneous Navier equation so that 
the normal stresses derived from It are 
zero at z = 0. That is, 

(A + 2/x)V(V.S 6 ) 


/* V X ( V X3 6 ) =0 


= a zz = Av • u + 2/x Vuzfdz | 

= a zp — jjl (dUg/dp + 3/xp/^Z ) Z 


where A and /x are the Lame coefficients. 
A general representation of w 6 satisfying 

Equation 2 is the Papkovich-Neuber 
representation (see Sokolnikoff, 1956, 
Ch. 6) 

5 & = (3 -^(* + %«?■?))//* (4) 

where a = (A + /*) / (A + 2(x) and 

A 2 $ v = v = 0, 1, 2, 3. (5) 

By symmetry the <£„ (and u b ) can be 
thought of as resulting from a point 
force at the image source point p = 0, 
z = — /i (see Fig. 102). Hence a useful 
representation for the § v satisfying 
Equation 5 is 

* = Y J B vl P l (cos x )/R' l +i (6) 

l = 

where R' 2 = /> 2 + ( 2 + ^) 2 an d cos x = 
(2 -f- h)/R f . We have assumed that 
$„ -> as i?' -> 00. The solution is ob- 
tained by substituting Equation 6 into 
Equation 4 and applying the boundary 
conditions Equation 3. One finds, then, 

3jj, -\- \ a 8V -> — ix SV cos v A 

$0 = — ; ; ^, , $ = 


A + fi 4nrR 

which results in a dilatancy of 

a R 





= V "U> 

= V 'u b 
V'?/(A + 2. i u) 

li 8V 1 — 3 cos 2 x 

A + fi 


R n 


At the surface (z = 0) for A = /* Equa- 
tion 8 reduces to 



2"[p 2 + * a ] 


{1 -3h 2 /( P 2 = h 2 )} 

We assume that there is a spherical 
dilatant region enclosing the eventual 
fault such that the radius h is one-half 
the fault length, and that the dilatation 
is uniform, centered at R = and of 
value 10 -5 . This is equivalent to a point 
source dilatancy of strength SV = 4/3 
7rh s X 10~ 5 . Equation 9 becomes 

o = - % x io- 5 

X [fc/V ? 2 + /i 2 ] 3 {1 - 3/i 2 /(p 2 + /i 2 )}. 


Figure 103 gives the threshold strains 

(6) based on Equation 10 as functions 
of epicentral distance {p) and fault 
length (2h) of the earthquake for which 
the dilatancy was a precursor. Also in- 
cluded are the magnitudes corresponding 
to fault lengths. For magnitudes below 
5 the relationship is the appropriate 
one for the Matsushiro region (m = 
3.47 + 1.90 logio 2h) and the relation- 
ships for magnitudes above 6% are due 
to Iida (1965). (That relationships ap- 
plicable to small and large earthquakes 
do not extrapolate into each other was 
pointed out by Chinnery, 1969.) 

From Fig. 103 for a fault length of 20 
km (nominally a magnitude 7 earth- 
quake) a dilatancy of 10~ 5 can be de- 
tected at about 40 km from the center of 
dilatancy by a strainmeter array of the 
type described above for a detection 
threshold of 10 -7 . 

The precursory dilatation detection 
threshold is a function of the instrument 
noise level, the amount of dilatancy, and 
the size of the dilatant region. For the 

5 / / 7 7.5 


Fault length ( km) 

Fig. 103. The maximum distance at which a precursory dilatant strain of 10" 5 would be 
detectable for noise thresholds of 10" 6 , 10~ 7 , and 10~ 8 for various fault lengths and magnitudes. 


instruments discussed, the threshold is Dorman, Borehole strainmeters, in 

about 10 -7 . However, this threshold Carnegie Inst. Wash. Year Book 69, 

could probably be improved significantly pp. 426-430, 19716. 

by installing the strainmeters at greater SokolnikofT, I. S., Mathematical Theory 

depths and thus reducing aquifer effects. of Elasticity, 2nd ed., McGraw-Hill, 

The amount of dilatancy is generally 1956. 
taken to be at least 10 -5 . The dilatant 

region must be at least the size of the Electrical Conductivity Studies 

earthquake but the large deviations in in South America 

Vp/Vb reported by various workers sug- L T AMHch and colleagues 

gest an even larger dilatant region. The . . . 

detection distances given in Fig. 103 Amon S the continuing geophysical 

therefore are probably minimum values. s ^ udies of the interaction of the Nazca 

plate and the South American continent 

T>p4 prpnC p« are those of the lateral variation in the 

radial structure of the electrical con- 
Cagniard, L., Reflection and Refraction ductivity. These studies were initiated a 
of Progressive Seismic Waves (trans- dozen years ago under the direction of 
lated and revised by E. A. Flinn and Dr. Ulrich Schmucker. With the collab- 
C. H. Hewitt Dix), McGraw-Hill, oration of several South American col- 
New York, 1962. leagues, Schmucker showed ( Year Book 
Chinnery, M. A., Earthquake magnitude 66, p. 12) the evidence for a most un- 
and source parameters, Bull. Seismol. usual structure whose minimum depth 
Soc. Am., 59, no. 5, 1969-1982, 1969. followed the eastern Andean cordillera 
Iida, K., Earthquake magnitude, earth- from northern Peru to central Bolivia, 
quake fault, and source dimensions, With continuing support from the Na- 
J. Earth Sci., Nagoya Univ., 13, 115- tional Science Foundation these collab- 
132, 1965. orative studies have continued in Peru 
Kasahara, A., A. Okada, M. Shibano, K. and Bolivia and in 1972 were extended 
Sasaki, and S. Matsumoto, Electro- to central Argentina and Chile. As this 
optical measurement of horizontal report is written a further extension of 
strains accumulating in the swarm the observations into the region of Co- 
earthquake area (3) , Bull. Earthquake lombia, studied seismically in Project 
Res. Inst., 45, 159-288, 1967. Narino during 1973, has been arranged. 
Nur, A., Dilatancy, pore fluids, and Argentina-central Chile with M. Cas- 
premonitory variations of t s /t p travel averde, J. R. Bannister, and L. Beach. 
times, Bull. Seismol. Soc. Am., 62, Last year's Report (Year Book 78, p. 
1217-1222, 1972. 1018) showed unusual areal distribu- 
Sacks, I. S., Strainmeters: long period tion of the H-Z transfer functions (TFs) 
noise, in Carnegie Inst. Wash. Year in this region. The long-period (one 
Book 73, pp. 1058-1060, 1974. hour) TFs were small and uniform over 
Sacks, I. S., and D. W. Evertson, A sensi- the whole region for two night-time 
tive borehole strain-rate meter, in variations. The short-period (15 min- 
Carnegie Inst. Wash. Year Book 68, ute) TFs showed a pattern of areal 
pp. 448-452, 1970. variation parallel and adjacent to the 
Sacks, I. S., S. Suyehiro, D. W. Evertson, coast line. Since the discontinuity of the 
and Y. Yamagishi, Sacks-Evertson oceanic conductivity is parallel to the 
strainmeter, its installation in Japan direction of magnetic field variation for 
and some preliminary results concern- those events, the explanation for the pat- 
ing strain steps, Pap. Meteorol. Geo- tern is not obvious. A third magnetic 
phys., 22, Nos. 3-4, 105-203, 1971a. storm with a large component normal to 
Sacks, I. S., D. W. Evertson, and L. M. the coast line now provides further in- 



70° W 


^Paso de las Aguas Negras 

Las Flores ' 


/ i 




. ® /# SANTIAGO 
Mario Pinto / 

30° S 


/ © * 

/ Cerro Negro 



+ Uspallato 





Stations recording 
June 17-18, I972,event 

— D-Z I hour period 
— D-Z 1/4 hour period 

Fig. 104. Areal variation of the D-Z transfer functions for a magnetic storm June 17-18, 
1972. The patterns for periods of 15 minutes and one hour are the same, but the absolute 
magnitude differs by a factor of 2. 

sight into the conductivity structure of 
the region. These results are shown in 
Fig. 104 with the contours for TFs of 
one-hour period nearly parallel to those 
of 15-minute period. These results indi- 
cate a conductivity anomaly with a N-S 
orientation in the ocean west of the ar- 
ray of stations. A similar result was ob- 
tained by Schmucker off the coast of 
southern California in 1960 (Schmucker, 
1970) . An uplift of the deep conductivity 
surface of the order of 100 km is implied 
by these results. 

Chile-Bolivia with J. R. Bannister, S. 
del Pozo, R. Salgueiro, and L. Beach. 
Measurements completed in Chile and 
Bolivia in previous years have been 
combined to produce the profile shown 
in Fig. 105. This figure shows the geo- 

graphic location of the stations studied 
and the profiles at both one-hour and 
one-quarter hour periods. There are only 
slight differences in the patterns, and the 
result implies a complex conductivity 
structure. There appears to be an uplift 
in the conductivity structure in the re- 
gion of, and at some angle to, the west- 
ern cordillera, and another east of Ta- 
rija. The angle is required since the 
cordillera is parallel to the magnetic 
field variation for the events studied. It 
is also possible that the complexity of 
the profile is due to the return current 
paths of the eddy currents generated in 
the anomaly present in southern Peru. 
In any case, there appears to be a pro- 
found difference in the results obtained 
from the stations in northern Chile and 



+ 0.2 r 



• I Hr 
X 1/4 Hr 



q Stations 

Fig. 105. Profile showing variation of the 
H-Z transfer functions for periods of 15 min- 
utes and one hour in the region of 22 °S lati- 
tude. The pattern for the two periods is iden- 
tical in form and magnitude. 

those 1000 km to the south, and the need 
for further observations to clarify the 
structure of the electrical conductivity in 
these regions is demonstrated. 

Several additional sites near the 
Cochabamba anomaly have been oc- 
cupied, and others will be added during 
the coming field season to outline better 
this unusual structure. 

Southern Peru with M. Casaverde, L. 
Tamayo, and A. Rodriguez. Data from 
six additional stations in southern Peru 
were obtained and, based on those data, 
additional sites in southern Peru will be 
occupied during the coming field season. 


Schmucker, U., Anomalies of geomag- 
netic variations in the southwestern 
United States, Bull. Scripps Inst. 
Oceanogr., 13, 1970. 

Local Seismic Net in Southern Peru 

A. Rodriguez and L. T. Aldrich 

The seismic network in southern Peru 
is currently recording at ah eight sta- 
tions with revised systems which permit 
operation at reduced power. In addition 
to operation of the station network, the 
Geophysical Institute of San Agustin 
National University is completing the 
reading of the 1969 data. It is hoped 
that the data reduction will be up-to- 
date by the end of the period of opera- 
tion of these stations. Because of previ- 
ous commitments, Dr. Asada will be 
unable to install his ocean-bottom seis- 
mometers until April 1976. It is our hope 
that the Arequipa net supported by the 
A. L. Day Fund and those supported by 
the H. O. Wood Fund will be able to 
operate concurrently, thereby providing 
the same kind of total coverage de- 
scribed last year. The new data will pro- 
vide information on earthquake distri- 
bution in both time and space over the 
last ten years. 

Cosmic Ray Research 
S. E. Forbush and L. Beach 

Cosmic-ray diurnal anisotropy 1937- 
1974 and the reversals of the sun's poloi- 
dal magnetic field. The Annual Report 
for 1971-1972 {Year Book 71) contained 
a discussion of the variation, from 1937 
to 1971, of the components of T 7 and W 
of the cosmic-ray diurnal anisotropy 
obtained from Carnegie Institution of 
Washington ionization chambers at 
Huancayo (Peru), Cheltenham-Freder- 
icksburg (U.S.A.), and Christchurch 
(New Zealand). Observations, first by 
Babcock (1959) and later by Howard 
(1972), at Mount Wilson Observatory 
since about 1953 show that the north 
polar magnetic field of the sun changed 
from positive to negative in 1958 and 
from negative to positive in 1971. These 
reversals coincided with the reversals in 
the component W of the cosmic-ray 
diurnal anisotropy in the asymptotic di- 



rection 128° east of the sun. This re- 
versal of W in 1971 now seems well es- 
tablished by the results, derived from 
observation through 1974, shown in Fig. 
106. The yearly values of the component 
V (not shown), with maximum in the 
asymptotic direction 90° east of the sun, 
are correlated with U (A) (r = 0.75) 
and exhibit a variation, analogous to 
that in sunspot numbers, with a range in 
each solar cycle of roughly 12%. W {A) 
is the yearly mean value of the south- 
ward horizontal magnetic field at the 
geomagnetic equator. U (A) is obtained 
quite simply from the empirical relation 
U (A) = - [H(A - Q) s - 0.68 
H{D — Q) s ]/Cos $, in which the first 
term on the right is the difference in the 
horizontal magnetic component for all 
days and the five magnetically quiet 
days at station s, and the second is the 
difference, disturbed minus quiet days. 
3> is the geomagnetic latitude of station 
s. The amplitude of W s , about 0.07%, is 
only about one-tenth of the standard 
deviation of hourly values at one sta- 
tion, which indicates one reason why the 
reliable determination of W s required 
data from three stations over several 

decades. This reduces the standard devi- 
ation of the departures of open circles 
from the wave W s in Fig. 106 to about 
0.018%. That for the departures of the 
open squares from W G {re, after 1966) 
is probably not much greater. 

When the cosmic-ray program was 
initiated in the early 1930's by the Car- 
negie Institution of Washington Com- 
mittee for Coordination of Cosmic 
Research, it was anticipated that by 
analysis of variations of cosmic-ray in- 
tensity from improved ionization cham- 
bers some evidence might be obtained to 
indicate the source of cosmic rays. From 
analyses of the data obtained in the pro- 
gram carried out at the Department of 
Terrestrial Magnetism, all the varia- 
tions (except the troublesome ones due 
to meteorological causes) are due to 
solar mechanisms. The last of these vari- 
ations to be reliably demonstrated is 
that discussed above and shown in Fig. 
106. This could have been derived only 
from an observational program of sev- 
eral decades' duration at several sta- 
tions. This variation must arise from 
some mechanism involving the long pe- 
riod variation (double the sunspot cycle 




5 - 

- -5 

1 1 



1 u 1 1 

o o 


- •/ . N 








/ \ \ 
/ \ \ 

/ o \ W G-4 




/o V 




-w s 



9 x 1 

' LEGEND \ 1 

, c 











20 YEARS \ 

AFTER 1966 

1 1 




1 1 1 





Fig. 106. Annual means of the diurnal anisotropy component in the asymptotic direction 
128° east of the sun fitted by a sine wave Ws from 1937 to 1965 and by an estimated wave Wg 
from 1966 to 1974. Measurements at Mount Wilson of the sun's north polar magnetic field 
showed that it reversed in 1958 and in 1971, as did Ws and W G . 



variation) in the sun's general magnetic 
field, although the solar field is much too 
weak to effect a magnetic cutoff at the 
earth (analogous to the geomagnetic 
cutoff near the earth due to the earth's 
magnetic field). 

Variability of cosmic-ray daily means 
over nearly four solar cycles. Figure 107 
shows the values, 1937-1974, of the 
yearly pooled standard deviation of 
cosmic-ray daily means from monthly 
means at Huancayo, Peru. Values are 
also shown with the five magnetically 
disturbed days of each month omitted. 
These standard deviations are generally 
only slightly less than those for all days, 
even though most of the large decreases 
of cosmic-ray intensity (Forbush effects) 
occur only during magnetic storms. Fig- 
ure 106 compares the variation of the 
yearly pooled standard deviations with 
that in yearly sunspot numbers. This 
shows the marked differences in different 
solar cycles of maximum sunspot num- 
bers and of the standard deviations. 
Similar variability is well known in geo- 
magnetic activity. 

Observations and reduction of data. 
Cosmic-ray ionization chambers were 
operated throughout the report year at 
Huancayo, Peru; at Fredericksburg, Vir- 

ginia, U.S.A.; and at Christchurch, New 
Zealand. Scalings and reduction of rec- 
ords have been maintained for these 
three stations. The reductions have been 
greatly facilitated by the use of the IBM 
1130 computer. 

Cooperation in operation of cosmic- 
ray meters. Appreciation is expressed to 
the U.S. Geological Survey and the U.S. 
Department of the Interior and its staff 
at the Fredericksburg Geomagnetic Cen- 
ter for efficient operation of the cosmic- 
ray meter there during the past report 
year, and to the Government of Peru 
and the Director and staff of the Insti- 
tute Geofisico del Peru for making 
cosmic-ray records from Huancayo 
available. Appreciation is also expressed 
to the Director and staff of the Geo- 
physical Observatory at Christchurch, 
New Zealand, for the excellent main- 
tenance of the equipment there and for 
the fine records obtained. 


Babcock, H. W., The sun's polar mag- 
netic field, Astrophys. J., 130, 364, 

Howard, R., Polar magnetic fields of the 
sun: 1960-1971, Solar Phys., 25, 5, 








- - SS NOS. 









Fig. 107. Yearly pooled standard deviations of cosmic-ray daily means from monthly means 
at Huancayo, Peru. 




Publications listed below can be obtained at 
no charge from the Department of Terrestrial 
Magnetism, 5241 Broad Branch Rd., N.W., 
Washington, D.C. 20015. When ordering, please 
give reprint number (s). Titles marked with 
asterisks are out of print. 

*4361 Aldrich, L. T., Report of the Treasurer, 
Trans. Am. Geophys. Union, 56, 307, 

Andersen, C. A., see Hofmann, A. W. 

*4362 Arnold, L. G., R. G. Seyler, L. Brown, 
T. I. Bonner, and E. Steiner, 2" 
threshold state in 8 Be, Phys. Rev. 
Lett., 32, 895-898, 1974. 

*4363 Arnold, L. G., R. G. Seyler, T. R. 
Donoghue, L. Brown, and U. Rohrer, 
Difference between polarization and 
analyzing power in the reaction 
3 H(p,n) 3 He, Phys. Rev. Lett., 32, 310- 
313, 1974. 

Assousa, G. E., see Thonnard, N., and 

Warner, J. W. 

Balick, B., see Warner, J. W. 

Beach, L., see Sacks, I. S. 

Bonner, T. I., see Arnold, L. G. 

4364 Brooks, C, and S. R. Hart, On the 

significance of komatiite, Geology, 
2, 107-110, 1974. 

Brooks, C., see Hart, S. R., and James, 

D. E. 

4365 Brown, L., and U. Rohrer, 3 H(p,n) 3 He 

reaction with polarized protons from 
threshold to 2.9 MeV, Nucl. Phys., 
A221, 325-332, 1974. 

*4366 Brown, L., N. Thonnard, and C. K. Ku- 
mar, The Z-dependence of vacancy 
production in asymmetric heavy ion 
collisions (abstr.), Bull. Am. Phys. 
Soc, 20, 639, 1975. 

Brown, L., see Arnold, L. G., Rohrer, U., 

and Thonnard, N. 

Cohen, G. N., see Costrejean, J.-M. 

Comaford, D., see Hofmann, A. W. 

4367 Costrejean, J.-M., N. Guiso, D. B. 

Cowie, G. N. Cohen, and P. Truffa- 
Bachi, The threonine-sensitive homo- 
serine dehydrogenase and asparto- 
kinase activities of Escherichia coli 
K 12, Eur. J. Biochem., 50, 431-435, 

4368 Cowie, D. B., Genetic interrelationships 

among the T-even bacteriophages, J. 
Franklin Inst., 298, 4, 1974. 
Cowie, D. B., see Costrejean, J.-M. 

4369 Cowley, A., R. Humphreys, B. Lynds, 

and V. Rubin, Report to the Council 
of the AAS from the Working Group 
on the Status of Women in Astron- 
omy — 1973, Bull. Am. Astron. Soc, 6, 
412-423, 1974. 





Craine, E. R., see Warner, J. W. 

Cuyubamba, A., see James, D. E. 

*4370 Dodson, M. H., Diffusion effects of a 
thermal pulse (abstr.), Trans. Am. 
Geophys. Union, 56, 472, 1975. 

Donoghue, T. R., see Arnold, L. G. 

— Duggan, J. L., see McDaniel, F. D. 
Erlank, A. J., see Hart, S. R. 
Fletcher, R. C, Wavelength selection in 
the folding of a single layer with 
power-law rheology, Am. J. Sci., 274, 
1029-1043, 1974. 
Fletcher, R. C, and A. W. Hofmann, 
Simple models of diffusion and com- 
bined diffusion-infiltration metasoma- 
tism, in Geochemical Transport and 
Kinetics, A. W. Hofmann, B. J. 
Giletti, H. S. Yoder, Jr., and R. A. 
Yund, eds., CIW Publ. 634, PP- 243- 
259, 1974. 
Fletcher, R. C, and R. H. McCallister, 
Spinoidal decomposition as a possible 
mechanism in clinopyroxene exsolu- 
tion (abstr.), Trans. Am. Geophys. 
Union, 55, 468, 1974. 

- Gardner, R. K., see McDaniel, F. D. 

- Giletti, B. J., see Hofmann, A. W. 

- Gladwin, M. T., see Stacey, F. D. 

- Goldich, S. S., see Hart, S. R. 
Grauert, B., Interpretation of Rb-Sr 

data of small whole-rock slabs and 
their minerals from the Manhattan 
schist, New York (abstr.), Interna- 
tional Meeting for Geochronology. 
Cosmochemistry, and Isotope Geol- 
ogy, Paris, France, August 26-31, 
Grauert, B., U-Pb systematics in hetero- 
geneous zircon populations from the 
Precambrian basement of the Mary- 
land Piedmont, Earth Planet. Sci. 
Lett., 23, 238-248, 1974. 
Grauert, B., M. G. Seitz, and G. Soptra- 
janova, Uranium and lead gain of 
detrital zircon studied by isotopic 
analyses and fission-track mapping, 
Earth Planet. Sci. Lett., 21, 389-399, 
Grauert, B., and M. E. Wagner, Age of 
the granulite-facies metamorphism of 
the Wilmington Complex, Delaware- 
Pennsylvania Piedmont, Am. J. Sci., 
275, 683-691, 1975. 

Grauert, B., see Hanny, R. 

Gray, T. J., see McDaniel, F. D. 

Guiso, N., see Costrejean, J.-M. 

*4378 Hanny, R., B. Grauert, and G. Soptra- 
janova, Paleozoic migmatites in an 
area of high-grade Miocene meta- 
morphism, Lepontinic gneiss com- 
plex, central Alps (abstr.), Interna- 
















tional Meeting for Geochronology, 
Cosmochemistry, and Isotope Geol- 
ogy, Paris, France, August 26-31, 

Hart, S. R., LIL-element geochemistry, 
Leg 34 basalts (abstr.), Geol. Soc. 
Am. Abstr. with Prog., 87th Annual 
Meeting, Miami Beach, Fla., p. 780, 
Nov. 18-20, 1974. 

Hart, S. R., and C. Brooks, Clino- 
pyroxene-matrix partitioning of K, 
Rb, Cs, Sr, and Ba, Geochim. Cosmo- 
chim. Acta, 38, 1799-1806, 1974. 

Hart, S. R., A. J. Erlank, and E. J. D. 
Kable, Sea floor basalt alteration: 
some chemical and Sr isotopic effects,, 
Contrib. Mineral. Petrol, U, 219-230, 

Hart, S. R., and S. S. Goldich, Most- 
ancient known rocks may be found in 
all earth's Precambrian shields, Geo- 
times, 20, 22-24, 1975. 

Hart, S. R., see Brooks, C., and White, 
W. M. 

Hastie, L. M., see Stacey, F. D. 

Hinthorne, J. R., see Hofmann, A. W. 

Hofmann, A. W., Diffusion of Sr and 
Ca in basalt melt (abstr.), Trans. Am. 
Geophys. Union, 56, 472, 1975. 

Hofmann, A. W., B. J. Giletti, J. R. 
Hinthorne, C. A. Andersen, and D. 
Comaford, Ion microprobe analysis of 
a potassium self-diffusion experiment 
in biotite, Earth Planet. Sci. Lett., 24, 
48-52, 1974. 

Hofmann, A. W., J. W. Mahoney, Jr., 
and B. J. Giletti, K-Ar and Rb-Sr 
data on detrital and post-depositional 
history of Pennsylvanian clay from 
Ohio and Pennsylvania, Geol. Soc. 
Am. Bull., 85, 639-644, 1974. 

Hofmann, A. W., see Fletcher, R. C. 

Humphreys, R., see Cowley, A. 

James, D. E., Initial Sr 8 7Sr 86 ratios in 
Mesozoic volcanic rocks of the cen- 
tral Andes (abstract), Trans. Am. 
Geophys. Union, 55, 475, 1974. 

James, D. E., C. Brooks, and A. Cuyu- 
bamba, Petrogenesis of central Andean 
Cenozoic magmas (abstr.), Trans. 
Am. Geophys. Union, 56, 474, 1975. 

Kable, E. J. D., see Hart, S. R. 

Kumar, C. K., see Brown, L. 

Lear, R. D., see McDaniel, F. D. 

■ Light, G. M., see McDaniel, F. D. 
Linde, A. T., Differences in radiated 

energy from suites of South American 
earthquakes at depths of 200 km and 
600 km (abstr.), Trans. Am. Geophys. 
Union, 56, 401, 1975. 

■ Linde, A. T., see Stacey, F. D. 

■ Lynds, B., see Cowley, A. 

■ McCallister, R. H., see Fletcher, R. C. 

*4389 McDaniel, F. D., T. J. Gray, R. K. 
Gardner, G. M. Light, J. L. Duggan, 
H. Van Rinsvelt, R. D. Lear, G. H. 
Pepper, and J. Nelson, K-shell X-ray 
production cross sections for 7 Li ions 
on selected elements Ti to Sb: 1.0 to 
5.0 MeV/amu (abstr.), Bull. Am. 
Phys. Soc, 20, 640, 1975. 

McKavanagh, B., see Stacey, F. D. 

Mahoney, Jr., J. W., see Hofmann, A. 


Nelson, J., see McDaniel, F. D. 

*4390 Okada, H., I. S. Sacks, and J. A. Snoke, 
Determination of the subducting 
lithosphere boundary by use of con- 
verted phases (abstr.), Trans. Am. 
Geophys. Union, 56, 394, 1975. 

Okada, H., see Sacks, I. S., and Snoke, 

J. A. 

Pepper, G. H., see McDaniel, F. D. 

*4391 Rohrer, U., and L. Brown, e Be(p,n) 9 B 
reaction with polarized protons from 
2.46 to 2.90 MeV (abstract), Bull. Am. 
Phys. Soc, 20, 693, 1975. 

Rohrer, U., see Arnold, L. G., and 

Brown, L. 

4392 Rubin, V. C, Two chains of interesting 

southern galaxies: NGC 7172-7173- 
7174-7176 and NGC 7201-7203-7204, 
Astrophys. J., 191, 645-651, 1974. 

4393 Rubin, V. C, D. Westpfahl, Jr., and 

M. A. Tuve, Second finding list of 
faint blue stars in the anticenter re- 
gion of the galaxy, Astron. J., 79, 
1406-1409 and 1501-1505, 1974. 

Rubin, V. C, see Cowley, A., and 

Warner, J. W. 

*4394 Sacks, I. S., Anomalous island arc 
asthenosphere and continental growth 
(abstr.), Trans. Am. Geophys. Union, 
56, 456, 1975. 

*4395 Sacks, I. S., and L. Beach, Lateral 
velocity variations at the base of the 
mantle and their correlation with 
gravity anomalies (abstr.), Trans. 
Am. Geophys. Union, 55, 350, 1974. 

4396 Sacks, I. S., and H. Okada, A compari- 

son of the anelasticity structure be- 
neath western South America and 
Japan, Phys. Earth Planet. Inter., 9, 
211-219, 1975. 

Sacks, I. S., see Okada, H., Snoke, J. A., 

and Suyehiro, S. 

Schilling, J.-G., see White, W. M. 

Seitz, M. G., see Grauert, B. 

Seyler, R. G., see Arnold, L. G. 

4397 Shimizu, N., An experimental study of 

the partitioning of K, Rb, Cs, Sr, and 
Ba between clinopyroxene and liquid 
at high pressures, Geochim. Cosmo- 
chim. Acta, 88, 1789-1798, 1974. 
*4398 Shimizu, N., Partitioning of K, Rb, Cs, 
Sr, and Ba between clinopyroxene 



and liquid at high pressures (abstr.), 
Trans. Am. Geophys. Union, 55, 473, 
4399 Shimizu, N., Rare earth elements in 
garnets and clinopyroxenes from 
garnet lherzolite nodules in kimber- 
lites, Earth Planet. Sci. Lett., 25, 26- 
32, 1975. 

*4400 Snoke, J. A., Archambeau's elastody- 
namical source-model solution and 
low-frequency spectral peaks in the 
far field displacement amplitude from 
earthquakes or explosions (abstr.), 
Trans. Am. Geophys. Union, 56, 401, 

*4401 Snoke, J. A., Earthquake radiation 
from an inhomogeneous medium with 
concentrated prestress (abstract), 10th 
Symposium on Mathematical Geo- 
physics, Intern. Union Geod. Geo- 
phys., Cambridge, England, Univer- 
sity of Cambridge, June 25- July 5, 

*4402 Snoke, J. A., Earthquake radiation spec- 
trum from an inhomogeneous medium 
with concentrated prestress (abstr.), 
Trans. Am. Geophys. Union, 55, 352, 
4403 Snoke, J. A., I. S. Sacks, and H. Okada, 
Empirical models for anomalous high- 
frequency arrivals from deep-focus 
earthquakes in South America, Geo- 
phys. J. R. Astron. Soc, 37, 133-139, 

*4404 Snoke, J. A., I. S. Sacks, and H. Okada, 
A model not requiring continuous 
lithosphere for anomalous high- 
frequency arrivals from deep-focus 
South American earthquakes (abstr.), 
10th Symposium on Mathematical 
Geophysics, Intern. Union Geod. Geo- 
phys., Cambridge, England, Univer- 
sity of Cambridge, June 25- July 5, 
4405 Snoke, J. A., I. S. Sacks, and H. Okada, 
A model not requiring continuous 
lithosphere for anomalous high- 
frequency arrivals from deep-focus 
South American earthquakes, Phys. 
Earth Planet. Interiors, 9, 199-206, 

Snoke, J. A., see Okada, H. 

Soptrajanova, G., see Grauert, B., and 

Hanny, R. 

*4406 Stacey, F. D., M. T. Gladwin, B. Mc- 
Kavanagh, A. T. Linde, and L. M. 
Hastie, Linearity of anelastic damp- 
ing: a conflict of evidence (abstr.), 
Trans. Am. Geophys. Union, 55, 1144- 
1145, 1974. 

Steiner, E., see Arnold, L. G. 

*4407 Suyehiro, S., and I. S. Sacks, Secular 
strain, strain steps and earthquake 
prediction criteria from observations 
in the Matsushiro region, central 
Honshu, Japan (abstr.), Trans. Am. 
Geophys. Union, 56, 400, 1975. 

*4408 Thonnard, N., L. Brown, G. E. Assousa, 
and H. A. Van Rinsvelt, Character- 
istic and continuum x-rays produced 
with potassium ions of a few MeV 
(abstr.), Bull. Am. Phys. Soc, 20, 639, 

Thonnard, N., see Brown, L. 

Truffa-Bachi, P., see Costrejean, J.-M. 

4409 Tuve, M. A., Early days of pulse radio 

at the Carnegie Institution, J. Atmos. 

Terr. Phys., 36, 2079-2083, 1974. 

Tuve, M. A., see Rubin, V. C. 

Van Rinsvelt, H. A., see McDaniel, F. 

D., and Thonnard, N. 
Wagner, M. E., see Grauert, B. 

4410 Warner, J. W., Narrow-band filter pho- 

tography of IC 4406, Publ. Astron. 
Soc. Pac, 86, 885-887, 1974. 

4411 Warner, J. W., On the stellar content 

and reddening in the nucleus of NGC 
5195, Astrophys. J., 190, 19-26, 1974. 

4412 Warner, J. W., G. E. Assousa, B. Balick, 

and E. R. Craine, Accurate positions 
and identifications for eleven Ohio 
survey sources, Publ. Astron. Soc. 
Pac, 87, 103-106, 1975. 

4413 Warner, J. W., and V. C. Rubin, Physi- 

cal conditions and structures in NGC 
7293, Astrophys. J., 198, 593-603, 1975. 

Westpfahl, D., Jr., see Rubin, V. C. 

*4414 White, W. M., J-G. Schilling, and S. R. 
Hart, Sr-isotope geochemistry of the 
Azores and the Mid-Atlantic Ridge: 
29°N to 60°N (abstr.), Trans. Am. 
Geophys. Union, 56, 471, 1975. 




L. Thomas Aldrich 1 
George E. Assousa 
Manuel N. Bass 2 
Ellis T. Bolton 3 
Louis Brown 
Dean B. Cowie 4 
W. Kent Ford, Jr. 
Stanley R. Hart 
Albrecht W. Hofmann 

Staff Members 

Bill H. Hoyer 
David E. James 
Alan T. Linde 
Nancy R. Rice 
Richard B. Roberts 
Vera C. Rubin 
I. Selwyn Sacks 
Norbert Thonnard 
Kenneth C. Turner 
George W. Wetherill, Director 5 

Distinguished Service Member of Carnegie Institution 

Merle A. Tuve 

Tom I. Bonner 6 
Nobumichi Shimizu 7 

Research Associates (Staff) 

J. Arthur Snoke 8 
Donald G. Wallace 9 

Research Associates (Nonresident) 

Mateo Casaverde, Instituto Geofisico del Leonidas Ocola, Instituto Geofisico del Peru, 

Peru, Lima, Peru Lima, Peru 

Cidambi K. Kumar, Howard University, Hiromu Okada, Hokkaido University, Sap- 
Washington, D.C. poro, Japan 

Shigeji Suyehiro, Japan Meteorological 
Agency, Tokyo, Japan 

Michael B. Davis 10 
George H. Pepper 11 

Carnegie Fellows (Postdoctoral) 

Charles J. Peterson 12 
J. Arthur Snoke 13 
John W. Warner III 14 

Carnegie Fellows (Predoctoral) 

John R. Bannister, University of Alberta, Arturo Cuyubamba, McGill University, 
Edmonton, Canada Montreal, Canada 

William M. White, University of Rhode 
Island, Kingston, Rhode Island 15 

1 Acting Director September 23, 1974, through 
March 31, 1975. 

2 Through March 15, 1975. 

3 Director, through September 23, 1974. 

* Institut Pasteur, Paris, France, from Janu- 
ary 20, 1975. 

5 From April 1,1975. 

8 Supported by U.S. Public Health Service 

7 Supported by National Science Foundation 
grant through August 31, 1974. 

8 Supported by National Science Foundation 
grant from April 1, 1975. 
"Through March 31, 1975. 

10 Through July 31, 1974. 

11 From September 1, 1974. 

12 From September 1, 1974. 

13 Through March 31, 1975. 
"Through September 30, 1974. 

15 From October 1, 1974, through February 
28, 1975. 



Visiting Investigators 
Christopher Brooks, Universite de Montreal, Martin H. Dodson, Leeds University, Leeds, 

Montreal, Canada 16 

England 17 
Raymond C. Fletcher 18 

Charles Doering 


Trainee Fellows 

Charles L. Bennett 20 
Antonio Flores 21 

Student Trainee 
David Westpfahl, Jr. 22 


C. Aguirre B., Universidad Mayor de San 
Andres, La Paz, Bolivia 

E. Arnal, Instituto Argentino de Radioas- 
tronomia, Villa Elisa, Argentina 

L. G. Arnold, Ohio State University, Colum- 
bus, Ohio 

E. Bajaja, Instituto Argentino de Radioas- 
tronomia, Villa Elisa, Argentina 

B. Balick, Lick Observatory, Board of 
Studies in Astronomy and Astrophysics, 
University of California, Santa Cruz, 

E. Berg, University of Hawaii, Honolulu, 

M. Bracamonte, Trujillo, Peru 

R. Cabre, S.J., Observatorio San Calixto, La 
Paz, Bolivia 

G. N. Cohen, Institut Pasteur, Paris, France 

R. Colomb, Instituto Argentino de Radio- 
astronomia, Villa Elisa, Argentina 

G. L. Davis, Geophysical Laboratory, Car- 
negie Institution of Washington, Washing- 
ton, D.C. 

S. del Pozo, Instituto Geofisico Boliviano, La 
Paz, Bolivia 

L. K. De Noyer, University of Illinois, 
Urbana, Illinois 

J. W. Erkes, State University of New York, 
Albany, New York 

A. J. Erlank, University of Capetown, Ron- 
debosch, South Africa 

D. W. Evertson, University of Texas at Aus- 
tin, Austin, Texas 

E. Filloy, Instituto Argentino de Radioas- 
tronomia, Villa Elisa, Argentina 

J. B. Flexner, University of Pennsylvania, 

Philadelphia, Pennsylvania 
L. B. Flexner, University of Pennsylvania, 

Philadelphia, Pennsylvania 
S. E. Forbush (retired), Department of Ter- 
restrial Magnetism, Carnegie Institution 

of Washington, Washington, D.C. 
C. Garavito, Planetario Destrital, Bogota, 

L. Gelinas, Genie Mineral, Ecole Polytech- 

nique, Montreal, Quebec, Canada 
A. A. Giesecke, Instituto Geofisico del Peru, 

Lima, Peru 
M. Gil, Instituto Argentino de Radioastro- 

nomia, Villa Elisa, Argentina 
M. Gordon, Instituto Argentino de Radioas- 

tronomia, Villa Elisa, Argentina 
J. A. Graham, Cerro Tololo Inter-American 

Observatory, La Serena, Chile 
C. Heiles, University of California, Berkeley, 


C. E. Helsley, University of Texas, Dallas, 

D. Huaco, Instituto Geofisico del Peru, 
Lima, Peru 

T. Krogh, Geophysical Laboratory, Carnegie 
Institution of Washington, Washington, 

I. Kushiro, Geophysical Laboratory, Car- 
negie Institution of Washington, Washing- 
ton, D.C. 

R. P. Meyer, University of Wisconsin, Madi- 
son, Wisconsin 

I. Mirabel, Instituto Argentino de Radioas- 
tronomia, Villa Elisa, Argentina 

16 From January 1, 1975. 

17 From September 1, 1974. 

18 Through August 31, 1974. 

19 From January 1, 1975. 

20 From May 15, 1975. 

21 From May 1,1975. 

22 Through September 30, 1974. 



R. Morras, Instituto Argentino de Radioas- 
tronomia, Villa Elisa, Argentina 

Y. Motoya, Kamikineusu Seismological Ob- 
servatory of Hokkaido University, Sap- 
poro, Japan 

G. Olafsson, Akureyri, Iceland 

E. Pimental, Cuzco University, Cuzco, Peru 

W. Poppel, Instituto Argentino de Radio- 
astronomia, Villa Elisa, Argentina 

J. E. Ramirez, S.J., Instituto Geofisico de los 
Andes Colombianos, Bogota, Colombia 

A. J. Ratkowski, New York University, New 
York, New York 

M. S. Roberts, National Radio Astronomy 
Observatory, Charlottesville, Virginia 

R. Rodriguez, Observatorio San Calixto, La 
Paz, Bolivia 

U. Rohrer, Swiss Institute for Nuclear Re- 
search, Villigen, Switzerland 

R. H. Sanders, National Radio Astronomy 
Observatory, Charlottesville, Virginia 

J. G. Schilling, Narragansett Marine Lab- 

oratory, University of Rhode Island, King- 
ston, Rhode Island 

R. G. Seyler, Ohio State University, Colum- 
bus, Ohio 

H. Sigtrygsson, Reykjavik, Iceland 

D. Simoni, Arequipa, Peru 

R. Stefansson, Reykjavik, Iceland 

L. Tamayo, Universidad Nacional de San 
Agustin, Arequipa, Peru 

J. L. Telleria, Planetario Destrital, Bogota, 

P. Truffa-Bachi, Institut Pasteur, Paris, 

H. A. Van Rinsvelt, University of Florida, 
Gainesville, Florida 

F. Volponi, Universidad Nacional de Cuyo, 
San Juan, Argentina 

R. L. Walter, Duke University, Durham, 
North Carolina 

Y. Yamagishi, Matsushiro Seismological Ob- 
servatory of Japan, Meteorological Agency, 
Matsushiro, Japan 

Supporting Staff 

Liselotte Beach, Research Assistant 
Kenneth D. Burrhus, Computer Systems En- 
Wayne Corley, Laboratory Helper, tem- 
Dorothy B. Dillin, Stenographer-Librarian 
John B. Doak, Electronics Research Spe- 
Clayton B. Doak, Grounds Maintenance, 

William N. Dove, Office Manager 
Everett T. Ecklund, Design Engineer 
Mark Feigensen, Laboratory Assistant, tem- 
Peter J. Fiekowsky, Laboratory Assistant, 

Laura A. Gazze, Research Assistant 23 
Gustavo Gonzalez, Laboratory Helper, tem- 
Mark Goudy, Grounds Maintenance, tem- 
Leo J. Haber, Carpenter and Maintenance 

Bennie Harris, Caretaker 

E. Kathleen Hill, Secretary 

Willis Kilgore, Jr., Caretaker 

Charles A. Little, Electronics Research Spe- 

Eric J. Lynch, Laboratory Helper, tem- 

Lilly Niu, Laboratory Assistant, temporary 

Niels M. Pedersen, Fiscal Officer 

Glenn R. Poe, Electronics Research Spe- 

Elliot M. Quade, Assistant Maintenance 

Carl M. Rinehart, Instrument Maker 

Judith H. Rogers, Typist 

Michael Seemann, Design Engineer-Mechani- 

E. Lynn Smith, Typist, Fiscal Assistant 

James E. Spicer, Assistant Maintenance 

Milton T. Taylor, Instrument Maker 

Neltje W. van de Velde, Research Assistant 

John M. Wilson, Laboratory Helper, tem- 

Through April 30, 1975. 

Hale Observatories 

Operated by Carnegie Institution of Washington 
and California Institute of Technology 

Pasadena, California 

Horace W. Babcock 

J. Beverley Oke 
Associate Director 


Horace W. Babcock, Chairman 

J. Beverley Oke, Vice-Chairman 

Jerome Kristian 

Robert B. Leigh ton 

Guido Munch 

Leonard Searle 

Arthur H. Vaughan, Jr. 

Harold Zirin 

Carnegie Institution of Washington Year Book 74, 1974-1975 


Introduction 307 

Observing Conditions 312 

Physics of the Sun 312 

The helium chromosphere 312 

Solar flares 314 

Synoptic observations 314 

Large-scale solar magnetic fields 314 

Large-scale velocity fields 315 

Systematics of solar magnetic field 

patterns 315 

Solar System Studies 316 

Infrared and UV observations of Venus . . 316 

Infrared observations of Jupiter 316 

Satellites of Jupiter 316 

Saturn 316 

Uranus 317 

Titan 317 

Io 317 

Stellar Spectroscopy 318 

Mercury stars 318 

Rotation of magnetic Ap stars 318 

Flare stars 319 

Stellar magnetism 319 

White dwarfs 319 

Studies of proper-motion stars 320 

AZ Cassiopeiae at the 1956-1957 

eclipse 322 

Stellar spectroscopy at 1 micron 322 

Spectrophotometric standards 322 

Eclipsing binary . 323 

Stellar Evolution 323 

Stellar Chromospheres 323 

Absolute magnitudes 323 

Chromospheric variations in main-sequence 

stars 324 

Millimeter and Submillimeter 

Photometry 324 

Observations at 350 microns 324 

1-mm observations of molecular 

clouds 325 

Observations of extragalactic objects at 

1 mm 325 

Interstellar Matter and Gaseous 

Nebulae 325 

Orion Nebula 325 

Orion Nebula at infrared wavelengths . . 326 
Infrared sources in molecular clouds . . .326 

Planetary nebulae 327 

Crab Pulsar 327 

Supernovae 327 

Supernova search 327 

Energy distribution and spectra 328 

The Galaxy 328 

Luminosity function of halo stars .... 328 

Galactic structure 328 

Galactic center 329 

The mass of the galaxy 331 

Galactic evolution 331 

Galaxies 332 

Ionization structure of gaseous nebulae 

in galaxies 332 

Compact galaxies 332 

Seyfert galaxies 332 

Velocity dispersions 333 

Dynamics of elliptical galaxies 333 

Infrared photometry of galaxies 334 

Metallicity of M31 and M32 334 

Globular clusters in M31 334 

Further observations of BL Lacertae . . . 334 

Radio source of 3C130 335 

Velocity fields in spiral galaxies 335 

Spiral structure in the old disks of 

galaxies 336 

Interacting galaxies 336 

Large Magellanic Cloud 337 

History of star formation in galaxies . . . 338 

Formation of galaxies 338 

Clusters of Galaxies 339 

Redshifts and magnitudes 339 

Redshift-magnitude diagram for a complete 
sample of nearby high-latitude Abell 

clusters 339 

Determination of the background light in 

the Coma Cluster 339 

Collisions of members of clusters .... 340 

Radio Sources 340 

Optical identification of radio sources . . . 340 

Radio sources near spirals 341 

Quasars and Quasi-Stellar Objects .... 342 

Quasar surveys 342 

Spectroscopic surveys 342 

Absorption lines 342 

Energy distributions 343 

Distribution of redshifts 343 

Variability 344 

Observational Cosmology 344 

Groups and clusters in the Southern 

Hemisphere 344 

Cluster redshifts 345 

Evolutionary effects 346 

Instrumentation 346 

PEPSIOS spectrometer 346 

SIT-Vidicon system for photometry .... 346 
Spectroscopy with the SIT-Vidicon . . . .347 
Wobbling secondary for infrared 

photometry 348 

Big Bear Solar Observatory (BBSO) 

instrumental program 348 

Las Campanas image-tube spectrographs . . 348 

Astroelectronics Laboratory 349 

Guest Investigators 349 

Photographic Laboratory 365 

2.5-Meter du Pont Telescope 366 

Dome and building 367 

Optics 367 

Microprocessor control system 367 

Auxiliary instruments 368 

Las Campanas Observatory 369 

Construction by contractor 369 

El Pino headquarters in La Serena .... 369 

Bibliography 370 

Staff and Organization 375 


The study of galaxies— their structure, 
dynamics, clustering, and evolution— is 
certainly one of the most active areas of 
research in astronomy today. This kind 
of study requires large optical telescopes 
(to say nothing of radio telescopes) with 
their indispensable light-gathering power, 
large image scale, and excellent optical 
performance. The new generation of 
optical telescopes currently coming into 
use around the world is already accel- 
erating research on galaxies to a very 
significant degree. Meanwhile, the pro- 
ductivity of the 5-meter Hale Telescope 
on Palomar Mountain is being enhanced 
by new image photometers and sky- 
subtraction spectrographs. 

A review of researcn topics reported 
upon for the past year at the Hale Ob- 
servatories shows that some twenty 
different projects or investigations re- 
lated to galaxies are in progress or have 
just been completed. A few of these 
have been selected for comment in the 
following paragraphs, but it is appro- 
priate first to refer to just two of many 
important milestones in the development 
of this subject. First was Hubble's 1936 
classification of shapes according to a 
bifurcated scheme with a sequence of 
forms branching from ellipticals into 
normal spirals and barred spirals. An- 
other milestone was the introduction in 
1962 by Eggen, Lynden-Bell, and Sand- 
age of the concept of galaxy formation 
by contraction and collapse of a primor- 
dial gas cloud to a plane, the difference 
in the formation history between ellip- 
tical and spiral galaxies depending on the 
relative rate of star formation in these 
systems. Of fundamental importance are 
the mutual viscosity of interacting parts 
of the gas cloud and the absence of 
strong interaction between stars and the 
cloud. It is clear that there is little evo- 
lutionary development along the sequence 
of forms. 

J. Richard Gott EI, of Caltech, and 
Thuan have further developed the con- 

cept of galaxy formation and differ- 
entiation already mentioned. They ana- 
lyzed the competing effects of dissipation 
and turbulent viscosity, using a model of 
cloud formation in the early protogalaxy. 
Numerical models, including both stars 
and gas, were used to simulate the col- 
lapse. The stars and gas collapsed to- 
gether until they reached the plane, 
whereupon the gas formed a disk and the 
stars continued in elliptical orbits to form 
a spheroidal halo. Some oscillations of 
the disk are expected before equilibrium 
is reached. Models may be used to relate 
galaxial parameters to the original size 
and angular momentum of the cloud. In 
general, if star formation in the early 
protogalaxy is essentially complete by 
the time of maximum collapse, an ellip- 
tical galaxy is formed. If a significant 
amount of gas remains at this stage, it 
will dissipate its energy to form a disk- 
like spiral. Observed properties of ellip- 
ticals and spirals imply that ellipticals 
formed out of larger perturbations of the 
primordial gas cloud than did spirals, 
giving the ellipticals greater densities and 
shorter collapse times. If early star 
formation is rapid and dependent upon 
the square of the density, ellipticals are 
expected to complete their star formation 
by the time of maximum collapse, while 
spirals are not. Protogalaxies arising 
from larger initial density fluctuations 
cluster more readily, accounting for the 
preferential occurrence of ellipticals in 
large clusters. 

Sargent collaborated with A. Boksen- 
berg and K. Shortridge of University 
College, London, in obtaining spectro- 
grams of ten elliptical galaxies with the 
5-meter Hale Telescope and Image-Pho- 
ton-Counting System. The purpose was 
to obtain the velocity dispersion for stars 
in each of these galaxies by analyzing 
Doppler broadening of lines. A typical 
result (for NGC 7626) is o = 285 ± 26 
km s -1 . 

C. P. Wilson also observed a group of 




ellipticals, obtaining spectrograms to be 
used for the determination of velocity 
curves of galaxial rotation and for veloc- 
ity dispersions. The SIT-Vidicon image 
photometer was used for part of this 
work, as well as for a study of the sur- 
face brightness in three colors. 

Schweizer continued his photometric 
investigations of broad spiral structures 
in underlying old disks of spirals. This 
was done in the near infrared, with a 
90-mm image tube, to show the enhanced 
surface-mass density in the old disk of 
M51, as contrasted to the narrower and 
younger spiral arms customarily observed 
in blue light. He also continued spectro- 
scopic studies of tidally interacting 
galaxies to find their relation to "normal" 

The strong, variable radio source 3C120 
was at one time thought to be a variable 
star in our Galaxy. Later, after it ap- 
peared that its spectrum is like that of a 
Seyfert galaxy and that its redshift is 
10,000 km s~\ the object was studied 
intensively because of its resemblance 
to a quasar. For some years there has 
been good evidence that some quasars 
occur in the nuclei of elliptical galaxies, 
so that when Arp reported in 1968 that 
there is a slight nebulous fringe around 
3C120, this was regarded as consistent 
with the foregoing. New photographs by 
Arp with the 4-meter telescope of the 
Kitt Peak National Observatory show a 
much greater extent to this surrounding 
nebulosity, and resolution of features 
within it. With other evidence, this 
has led him to the conclusion that 3C120 
is a spiral or disturbed spiral galaxy 
whose nucleus contains the active, com- 
pact, and variable source of energy. 

The best hope for advancing know- 
ledge of the chemical composition and 
chemical evolution of galaxies lies in the 
interpretation of the spectra of the re- 
gions of ionized interstellar gas within 
them. Searle has recently surveyed the 
strengths of emission lines that are 
probes of ionization structure in a num- 
ber of H II regions in M31, M101, and 
M51. He has carried out a systematic 

survey of the behavior of the infrared 
lines of [S III] and is collaborating with 
Dr. G. Shields of the University of Texas 
in the interpretation of the data. The 
observed behavior of the [S III] lines is 
not in accord with the predictions of 
the simple ionization-structure models 
that have so far been used to interpret 
the spectra of H II regions. Unexpectedly, 
large density fluctuations within the 
emitting regions are indicated. It is 
already clear that the low excitation 
that is characteristic of regions situated 
in the inner spiral arms of galaxies is 
not simply a consequence of an enhanced 
oxygen abundance, as had earlier seemed 

Zinn, collaborating with Dr. John 
Norris of the Mount Stromlo Observ- 
atory, has investigated the stellar popu- 
lations of dwarf spheroidal galaxies. 
Color-magnitude diagrams for such 
galaxies and for globular clusters in our 
Galaxy long ago suggested similarities 
in star formation for the two groups. But 
variable stars in the dwarf spheroidal 
galaxies and in the globular clusters do 
not obey the same period-luminosity 
law. Norris and Zinn believe that the 
period-luminosity law, the evolutionary 
status of the variables, and the color- 
magnitude diagrams can be explained if 
there were two epochs of star formation 
in each of the dwarf galaxies, but with 
little enrichment of metal content in the 
intervening time. They suggest that 
either supernova explosions were unable 
to drive all the gas from these systems 
or that gas was accreted some time after 
the first epoch. 

Infrared observations of galaxies in the 
1- to 3-micron range were continued by 
Jay Frogel, E. Persson, and M. Aaronson 
of Harvard College Observatory in collab- 
oration with Becklin, Neugebauer, and 
Matthews. They made use of the lumi- 
nosity-sensitive CO index at 2.3 microns. 
Initial results showed that the radiation 
from the nuclei of elliptical galaxies is 
dominated by the light of giant stars. 

Becklin and Neugebauer have made 
new high-resolution maps of the region 



of the center of the Galaxy at wave- 
lengths of 2.2, 10, and 20 microns. They 
conclude that over half of the 2.2-micron 
radiation from the central 2 parsecs of 
the Galaxy comes from discrete unre- 
solved sources. These objects could be 
luminous individual late-type stars or 
very compact clusters of stars. At 10 
microns, nine discrete sources of radi- 
ation were detected in the 2 parsec cen- 
tral region. 

Observations bearing on the distribu- 
tion of metallicity in the Andromeda 
Spiral, M31, were obtained by Dr. M. 
Schwarzschild of Princeton University in 
collaboration with Oke. They used the 
multichannel spectrometer on the Hale 
Telescope to obtain energy distributions 
at several locations in the central bulge 
of M31. The results can be interpreted 
as an increase of the heavy element 
abundance toward the nucleus of the 
spiral. Beyond a radius of 2', no further 
abundance changes were observed; but, 
at 7'.5 and 10' from the center of M31 
along the major axis, there is evidence 
for young hot stars. 

Van der Kruit reported on the reduc- 
tion of spectrograms of three spiral 
galaxies he had obtained with the image- 
tube spectrograph of the Hale Telescope. 
Detailed velocity fields were constructed. 
Noncircular motions in NGC 4736 were 
of particular interest. An excellent cor- 
relation was found between the mag- 
nitude of such motions and the bright- 
ness of the radio continuum emission. 
In NGC 3310, a systematic pattern of 
strong noncircular motions was found 
associated with the bright spiral arms. 

Gunn and B. M. Tinsley of Yale Uni- 
versity are continuing their program to 
synthesize the stellar populations of 
elliptical galaxies in an effort to deter- 
mine the evolutionary corrections to the 
Hubble diagram. 

Sandage completed analysis of red- 
shifts and photometry of selected groups 
and clusters of galaxies in the Southern 
Hemisphere. Two results of particular 
interest emerged. When the redshifts 
were plotted against magnitude to con- 

struct the Hubble diagram for these 
groups, it was found that the residuals 
in absolute magnitude were only mildly 
correlated with population of the group. 
When aggregates of galaxies having 
populations in the range 20 to 300 are 
examined, it is found that the mean 
absolute magnitude of the first-ranked 
galaxy varies by only about 0.15 mag. 
This suggests that the absolute mag- 
nitude of the first-ranked galaxy is deter- 
mined by some physical upper limit in 
the galaxy formation process that is 
independent of the statistics relating to 
other members of the cluster. 

The new data from the Southern Hemi- 
sphere, when analyzed in terms of veloc- 
ity differences and combined with exist- 
ing northern data, failed to show any 
systematic velocity perturbation in the 
local velocity field. New and closer limits 
can be placed on the perturbation of 
the local Hubble rate H for galaxies 
having velocities smaller than 3000 km 
s _1 toward or away from the Coma- Virgo 
complex. A lack of perturbation in the 
velocity field at this level, in the presence 
of the density contrast caused by the 
clumping of galaxies in the Virgo group, 
shows that the effect of gravity on the 
underlying linear velocity flow is small. 
This suggests that the kinetic energy in 
the expansion field is large compared to 
the gravitational potential energy of 
the density contrast, and hence that the 
deceleration parameter q is very small, 
in agreement with the results reported 
last year. 

Gunn and Oke have continued their 
separate program to discover faint clus- 
ters of galaxies and to measure the red- 
shifts and absolute spectral-energy dis- 
tributions of the brightest members of 
these clusters. The first aim is to produce 
a list of clusters found under well-con- 
trolled conditions within a definite limit 
of apparent magnitude. Hundreds of 
clusters were identified on plates ob- 
tained in the course of an initial survey 
with the 1.2 -meter Schmidt telescope. 
Spectrophotometric observations of a 
few show that redshifts are typically 



between 0.20 and 0.35. A deeper survey, 
conducted with a 90-mm image tube at 
the prime focus of the 5-meter telescope, 
has produced a sample of about 30 very 
faint clusters of galaxies that typically 
have redshifts larger than 0.40. 

Halo stars of our own Galaxy have a 
roughly spherical distribution about 
the galactic center. Using a complete 
sample of stars of large proper motion 
with known trigonometric parallaxes, 
Schmidt derived the local luminosity 
function of high-velocity stars that belong 
to the halo. The luminosity function was 
deduced by using a kinematical criterion 
based on space velocities of metal-poor 
RR Lyrae variables. Schmidt found the 
local mass density of halo stars to be 
1.7 x 10" 4 3TI pc -3 ; he calculated that 
the halo contributes about 6% of the 
mass of the Galaxy interior to the loca- 
tion of the sun. 

The foregoing brief reports constitute 
a sampling of galaxial research at the 
Observatories. References to some other 
fields of current interest are provided 

The mode of excitation of helium in 
the sun's atmosphere has long been of 
interest. Because helium is difficult to 
excite, it was thought to be excited either 
by coronal ultraviolet radiation or by 
high chromospheric temperatures. Space 
observations showed that the resonance 
lines of He II showed minima in the 
regions known as coronal holes, where 
the corona is less dense. With the 65-cm 
vacuum telescope at Big Bear, Zirin 
found that the D3 helium emission line 
also disappears in regions of coronal 
holes and has a sharp lower limit about 
1000 km above the photosphere. These 
two facts were combined to produce a 
model of the excitation of chromospheric 
helium by back-radiation from the co- 
rona. According to this model, the cor- 
onal radiation shortward of 500 A ionizes 
helium to the depth of its penetration, 
producing the observed dark band below 
the region of optical depth unity. The 
helium lines seen are produced by recom- 
bination from the ionized state for both 

neutral and ionized helium. Calculations 
based on this model gave good agreement 
with the intensities of D3 and other 
helium lines, including those in the 
ultraviolet. Two corollaries follow: 

The solar helium abundance cannot be 
regarded as well determined because the 
helium intensity depends on the variable 
ultraviolet from above and not on the 
amount of helium. The observed helium- 
line absorption in late-type stars may 
also be attributed to coronal back-radi- 

Preston, Vaughan, and former Car- 
negie Fellow Richard E. White have 
completed their investigation of Hg 
isotope anomalies in the mercury stars. 
Wavelengths of Hg II A 3984 in 30 stars 
are distributed rather uniformly from 
the value for the terrestrial mix of iso- 
topes to a value that implies a prepon- 
derance of 204 Hg. The wavelengths are 
loosely correlated with effective temper- 
ature. Relative isotopic abundances 
derived from partially resolved profiles 
of A 3984 in t Coronae Borealis, x Lupi, 
and HR 4072 suggest that simple mass- 
dependent fractionation has occurred 
in all three stars. This indicates that an 
isotope separation process rather than 
exotic nucleosynthesis is responsible for 
the isotope anomalies. Accordingly, a 
scheme has been devised whereby iso- 
topic compositions can be inferred from 
a comparison of stellar wavelengths and 
equivalent widths of A 3984 with those 
calculated for a family of fractionated 
isotopic mixes. For one star, 46 Aquilae, 
Preston reports a wavelength for the 
Hg II line that corresponds to pure 204 Hg. 

Greenstein has continued the color- 
imetric search for white dwarfs, espe- 
cially those with late-type spectra and 
colors, using the multichannel spectro- 
photometer on the Hale Telescope. The 
rarity of yellow and red degenerate 
stars continues to contradict the simple 
theory of cooling of the core of a degen- 
erate object. Although numerous objects 
have been found cooler than 8000 °K, 
none have been found cooler than about 
4500 °K (EG 202). A few have relatively 



strong metallic lines, but at wavelengths 
longer than A 5500 they prove to be 
relatively blue objects. The reasonable 
possibilities seem to be that (1) the cool- 
ing becomes rapid below effective tem- 
perature 5500 °K and the stars disappear 
into the black dwarf stage, or (2) some 
physical mechanism delays cooling of old 
white dwarfs. 

In cooperation with Dr. A. Boksenberg 
of University College, London, Green- 
stein used the image-photon-counting 
system in a study of spectra to show 
that all suspected late-type white dwarfs 
previously classified as DM are weak- 
lined subdwarf M stars, i.e., very metal 
poor. No degenerate star with TiO bands 
was found. 

The magnetic degenerate stars provide 
the most unusual spectra; study of sev- 
eral such stars suspected of showing 
variable circular polarization demon- 
strated that the broad absorption fea- 
tures found and not yet identified do not 
move substantially; i.e., quadratic Zee- 
man effect does not grossly falsify their 
wavelength. The periodic variable G195 
-19, however, shows a trace of spec- 
trum variability in its broad features, 
changing in location or intensity. The 
broad and somewhat wavelike distortion 
of these spectra approaches the limit 
of detectability. Magnetic fields of up to 
10 8 gauss are suspected. 

0. C. Wilson has completed his work 
on absolute magnitudes of stars obtained 
from measures of the width of chromo- 
spheric emission of the K-line of Ca II, 
and the material is being prepared for 
publication. The table of results contains 
useful information on luminosities for 
about 700 objects. These stars are of 
spectral types G, K, and M; subgiants 
and supergiants are included. One of the 
principal results is that the lower bound- 
ary of the color-magnitude diagram 
plotted from the tabular data conforms 
closely to the normal points found earlier 
by Sandage for NGC 188. This indicates 
that all of the nearby field stars listed 
are about the same age as the cluster. 
There is no real evidence for the presence 

of an older population. The K-line data 
are very homogeneous, since they depend 
upon measures by one individual on 
spectrograms of the same dispersion. It 
is, therefore, of interest to compare 
them to the results from trigonometric 
parallaxes. This was done by plotting the 
published parallaxes against the moduli 
from the K-line measures to see how they 
fit the curve m - M - -5-5 log n. 
Stars with galactic latitude b > 25° 
provide an excellent fit to the curve, but 
those with b < 25° are shifted in the 
sense that their trigonometric parallaxes 
are too small. Wilson concludes that 
there are most probably errors in the 

Sargent, with R. F. Carswell and A. 
Boksenberg of University College, Lon- 
don, applied Boksenberg's image-photon- 
counting system in observations of the 
17th magnitude quasar PHL 957 at the 
coude spectrograph of the Hale Telescope. 
This quasar has z em = 2.69 and a very 
rich absorption spectrum. In all, 284 
absorption lines were measured, and 
intensive effort was put into a search 
for redshifts, using computer techniques 
developed for the analysis of Pks 0237 
-23 last year. However, only one ab- 
sorption redshift, 2 a bs = 2.30 is regarded 
as definite, and this accounts for only 
about 5% of the lines. 

Westphal and Kristian have continued 
to refine and test the application of SIT 
television camera tubes for two-dimen- 
sional (image) photometry. The level of 
photometric precision obtained at the 
telescope is close to the 1% limit expected 
from laboratory tests of the system. 
Over a range of 10 magnitudes, from 
12 to 22, it is possible with the SIT- 
Vidicon to reproduce with highly satis- 
factory precision a broad-band sequence 
established with photomultipliers. The 
limits are not intrinsic, but at magnitude 
22 the signal is completely dominated 
by the sky. SIT measurements should be 
accurate to fainter levels, although fainter 
photoelectric standards are not available. 
In a 10-min exposure, stars are measur- 
able fainter than V = 24 mag, with pro- 



gressively lower precision attributable 
to photon statistics. Currently, the 1% 
limit for precision of the SIT system is 
due primarily to seemingly random 
variations of the order of 0.5% in the 

The same SIT system used in image 
photometry has been employed by West- 
phal and Kristian for several programs 
of low- and high-resolution spectroscopy. 
When used with a small spectrograph, 
there are two important features of the 
SIT detector that make it particularly 
suitable for faint objects. The repro- 
ducibility of the transfer function allows 
accurate sky subtraction, while the two- 
dimensional format permits the spectra 
of the sky and object-plus-sky to be 
recorded simultaneously. The sky-sub- 
traction process functions well and pro- 
duces difference spectra that are limited 
essentially by photon noise. The SIT- 
Vidicon has also been applied success- 
fully as a detector at the coude spectro- 
graph of the Mount Wilson 2.5-meter 

telescope, where it has been used by 
Brucato to study stellar line profiles and 
by C. P. Wilson to measure velocity 
dispersions and rotations of the central 
regions of galaxies. Cooling of the tube 
to -90°C makes possible exposure 
times of several hours. 

An experimental grating spectrograph 
equipped with a cooled integrating SIT- 
Vidicon camera is nearing completion 
under Gunn's direction. The system is 
being interfaced with a PDP-11/40 mini- 
computer to give real time capability for 
addition, averaging, reduction, and 
display of spectra. 

A new f/60 infrared photometer has 
been put into operation on the Mount 
Wilson 1.5-meter telescope. Signal modu- 
lation is obtained by wobbling the Casse- 
grain secondary of the telescope. The 
system provides a very satisfactory level 
of performance for the study of infrared 
radiation from molecular clouds, H II 
regions, and external galaxies. 


The 2.5-meter Hooker telescope at 
Mount Wilson was used for observations 
on 240 complete nights and 36 partial 
nights for a total of 2579 hours. The 
telescope was out of service for six nights 
to permit replacement of main shaft 
bearings in the clock drive. The 1.5-meter 
telescope was used for 144 complete 
nights and 79 partial nights for a total 
of 1700 observing hours. The primary 
mirror was removed and taken to Palomar 
for realuminizing in July 1974. Rainfall 
for the year was near normal at 914 mm, 
and total snowfall was 1473 mm. 

The 5-meter Hale Telescope at Palomar 
Mountain was used for a total of 3473.3 
hours, of which 2892.8 were nighttime 
hours, as shown in Table 1. The differ- 
ence represents twilight time that was 
used for infrared observations not re- 
quiring a dark sky. Total precipitation 
at Palomar was 730.5 mm, with 1366.5 
mm of snow. The maximum temperature 
was 32.2 °C in July; minimum was -8.3° 
C in December. 

Public visitors at the Palomar Observ- 
atory numbered 140,000 for the year. 


The Helium Chromosphere 

For many years the mode of excitation 
of helium in the sun's atmosphere has 

difficult to excite under normal temper- 
atures, it was considered that it was 
excited either by coronal ultraviolet 
radiation or by high chromospheric 

been of great interest. Because helium is temperatures.